Occasional papers of the California Academy of Sciences

No. 149

OCCASIONAL PAPERS
OF THE
CALIFORNIA ACADEMY OF SCIENCES

September 21, 2000

EMBIA

Contributions to the Biosystematics
of the
Insect Order Embiidina

Part 1
Origin, Relationships and Integumental
Anatomy of the Insect Order Embiidina

Part 2
A Review of the Biology of Embiidina

By
Edward S. Ross

Published by the California Academy of Sciences
San Francisco, California

erp
rs

ver

EMBIA
Contributions to the Biosystematics of the
Insect Order Embiidina

Male Clothoda n. sp., (in alcohol, body length 20 mm), related to C. longicauda Ross, of Peru, second only
to C. nobilis (Gerst) as the most plesiomorphic species of the order. Plesiomorphic features include: large
size, complex wing venation and almost completely symmetrical terminalia, including exceptionally long
cerci. Specimen from upper Napo River region of Ecuador (see Figs. 39A and 40A).

OCCASIONAL PAPERS
OR EE
CALIFORNIA ACADEMY OF SCIENCES

No. 149

EMBIA

Contributions to the Biosystematics
of the
Insect Order Embiidina

Part 1
Origin, Relationships and Integumental
Anatomy of the Insect Order Embiidina
Part 2

A Review of the Biology of Embiidina

By
Edward S. Ross

Published by the California Academy of Sciences
San Francisco, California

SCIENTIFIC PUBLICATIONS

Alan E. Leviton, Editor
Katie Martin, Managing editor

© 2000 California Academy of Sciences, Golden Gate Park, San Francisco, California 94118
All rights reserved. No part of this publication may be reproduced or transmitted in any form or
by any means, electronic or mechanical, including photocopying, recording, or any information

storage or retrieval system, without permission in writing from the publisher.

ISSN 0068-5461

Table of Contents

Part 1. Origin, Relationships and Integumental Anatomy of the Insect Order Embiidina

| EIA E:T eL aa eta ec es cea ar er NCOP A cee, ee Ue tere See aa yrs I NA 1
Acknowledgments! win tet yee vccseceree tree rail terck ae choi ec dere teaiy Revo) Site eRe see sy cries i sieaateeaeree l
Origin and Relationships
SLITS Tinta Sha Oeei eg, Stars ec her ieee seen ba snoc ene oes een Mr tress Cement OccLas aero Gicyre eR ere SOS ORD CEE Fo 3
Relationships yessciccisscasers emeten ss ists sev bere ile ses sera cahetehs Mec caamew eRe tn Perennials SVs Ree SEO ACE 3
ON Neeiera ene oe Sees Sean gO cn mo means Oo oaaT Coe one oot eatae kd teos ron 3
Rossiltrecord eer mat pets ante tee eres Se Gi DO Ane aT PE ISIE Ce 4
retertianyafOSsilSimccrane cra tee cot ahs epee SI Ran a a pane Pee eee 5
ANS MET AY ROSS 35.5 a gam eeadien cm cotb\o 0.21 BS RIM ere rene Can Se Ce Eon an OU mia atoart ore aA 6
Integumental Anatomy
Wnt AUC EOE R Ae aye estes te ee, Hert eese ey Teak ren Oa ces Sachs eatery Stel anc Honea Sa ects aia ecieue creat a Ry en coer, eg 7
Meth OdSiecieroeiqe cen creia setndea pena Sh geiatasalurunnenicee ainda tamuvin ale cae nls neacon cure ei yee Staion i erage 7
Ve (SPz a os cya oa’) akc ec acy ee ee Ce era Ch REE aOR ea a hl enc lees, en nti eG 8
1 BS etree ech crth Gere OLe ec LSE CREO roe ten ara et ee cohen State Nema aU my rte aaron ome 11
PATI LETITIA aaa fe cored tees neare nape recor fecad cece ad eoeeneeen ted ate eta ciaauemm teed hel cree eer erate teed era ee 1]
Mouth partseeersr ners se garcacnsecueroctcasacresnsnet ae resie eee mame RT EU PU MPR ener ee or Ser eecmere tera 12
(Qin inb Seite ae orator eee een SCG Ta cnr et een Gracia Miokcicia Hae anni Gurnee eames ao ery og Z 13
RrOthoraxss ects, sence cere nieheyn se ccuc ne de ereser ose erie ops gee An RO a nol cH7e ea berate ean ceed eae 15
Bteroth onaxofprmial Siar nce csc scnshe st svete anes pencin cene eae ecte oh et coe var a tat nca omer aiearan deepaeoe tease era ra 16
Meso=!and!metathoraxofi females: ci.e eccecicsrceis ntiercree eneionc eet se ele tees ste tose cute peneianyee 18
EC BS peers ree erie tte nee atau vurecners, ss lcuauctaicey rah cue ee Minh Cecraeh En tah duepadtn ea etch tie ceed wegeaces tare Mea Et cane 21
rothoracicile psc sw se mecca trey est rcs recat eh ieee aero nia en nee 21
INaturero fier binds il katte ce eran tsa teter te iceecegec ast sve. cctentheiteususlicus eutratees doe Rie bate eeu heb eeeteueteee ater 23
Mesothoracicslegstie arr ctws co ctotene crete es eee ier ckere eeegsre tare aren erin oyna iets teks env ue eW 23
Metathoracichle ose ecat thn sian erosersucsieccns mcrae eee one noreay cries eer rs ar ae ope eae at cance en ae 24
AWAIT SSW chpeiae eter cats see eey Asm ofc yeitsy ve abe EGU ea ss ete cease tzesie ova aus ya ited curate) Aucucwtre| Maal yeu cee cuey Runt Sve tee 25
Vem ALLO Mae cree eh Ua rare ee mts yee sMeete st 3) ces aallsuahayeccxctn, cot te see east raat snannuh meeeeae cy ae mea 27
Win Sani simientatl ontersyeeecen tmeene setae ce Steve satires aalnuactye peti era ca neueveleay vic reenter eeore + Sues 30
WinggexpansionsfollowinprecdysiSumecmeacse cued icisctereaae ecieeadiane mic cu reais cacecuelererete tee 33
Whinggarticu lagiony cece ca ctavoccrsvcteyckcerane tit cntassas caret Teo rie ei cae reenact ue aetare aa eee tee 33
FE XA O MIO FOWAT OS 5. e coy Shairerciese she actin es cx citrate aoe tay ea oe ST aE eco ae eos aes eae 34
el hts Sreeusee epee weiss ate getters yen teitaecei inves a-oahays Sane team pce MUA tae NEWS eda eA Sad tadin eee ee 34
ReductionyandteliminationofawinkSeeermet ioe rrr te cree ea renee knee exe aarti: 34
JN oXOFOSO NEL ee cucy eEeL ROAR IAC ONE CHET) CORE ce GN eRe OEE oro MC TE OIE Oey Oss aikaed. 6 38
Externaligenitaliavofefemalesinarimreascrs ger ce cencntc re nenetieea teres se crict aver ek neat titers ewe noueusn ace het 39
Internaligenitaliatofimalesta vce thee ne ots orcs cnn se ps eh eae espe onsgtroenecusTa eee tae Reo 39
External:genitaliavofimales\ (je Merminaliag) eee eeepc tces ease ceee cee es yaks sloystor te) sae esi-weiccte ee 40
Anomalous:maleiterminaliaas secre caste cers ee excel eieucaeoreycueio cei lee se atsionerey cies aes beens 46
PortfoliovofmaletterminaliainsEmbiidina’ ¢......3 pc. eee were ¢ ee geen yew os eens snes © eon 47
ST eLAIT Ca CH tele anit no ee. Ot Sido Vena AEA SR CRT oes cnr ame AC iner bins aaeteras Seca me 52

SUMIMAN Ye seapere cresy car heire cope asesce ari useers en neuer spebstnn cisvies a cere eCtotina sue teumuse ya tacebeh site Ook UP Aints cue eeea hin bakes orate ]
MIGHTY oben oc eroo ae ee OOS Oct MOOR Oe DDE TOA GOCE ERE RAGE MERA TA GG Mane dé pants 1
Genetalibi ology mirtrcense tres ce teae haere sctey each els wera ee castro alts nec eat: RUPEE PPE CT ee I
SOcialyDeh avin ecw ct scare tee eager sey We pose seca Ser atte negara avo ener ean eM nape te Pee once ees eke Ses ook easement sae RTS) pene 8

Golomotiadultimales: cs 2 eck. ssi eoees te ct) A SIE I ree 10

IM ENGLIVER 6 Gree, Cpe ero ene: ROC e CE ac ets kote aan er eae IIA MESA ei iol SrolcimnnOInOIe GO GE'S'6:0'0'0:0.06.0 11
Eggsiandithein protection’ ake ts 22h, Sets a aie estate el nln cons UAT Sees hsv et qe ae a Un 13
Mevelopmientyrec.ctebescnes ci essycevs aioe Gases ceens aie Stele Oo) Seth alel eat oh ETP COC TORO eee ROS 15
EEX PANS OMVOLAWINGS caccveces vrs faane «aco opm eisy evens sien. eaayeis a won auecebiers Inthe Aerie Cae eeRere Te aoe 16
RarthenOpgenesisyac ri seca. costs tone eects coe eee ae cece eee ate ae ER Ee, Seale SiO cece eno 17
1 DY cnc eon acta ee aoe Bn eed art a a ee ree raters Cera Gt SG Sicna\G-o\d-0'0 010-0 0 18
IMovementent: sate choccy it eens. 2 ean lee nap een en eo Dilan Se eee 20
Habitats
MropicallEversreenvkOrests) |< crn eae ye oehs cane che rs renee cine Oro. eee yee Re eee 21
MropicallCloudiFOrests shy se asst cesses sve Rade cee actus ern cartes hecat os ak nce meee eee aR 22
Seasonally-Dry‘Grassy Woodland 3.22 dic oats oh een eee eens Genes cen sree 22
Semi-AridOpen'Grasslands' <0. occas nese cetceso cathe y ces cee svete the cierto eet eee eae ere 24
eS ertvAT CAS oy. cece ese ceses tos eel efay's)n; Oliov a eda cal unbedicvele level chosen niinierse neu iclGtanereratate CLR eee ee 25
HumaniHMabitats: aes 5ee nsec org ects esahe teva ees eLevoimcatie Se. ninieita se even Tae eye ATOR ee RE ee 25
Geooraphiciran ge clin. oe .cde the mays ceetee Cus ciate e ty chaleos ekeseheustecaiele eitsey cease tian oie teens cee eee 28
Ecologicaliran ge: 22.2. tintenty: tee Ptuers ease iatevtaonsuccettos) saehe eee deoetianeteve se eeriana Sie ele niet Pee eee ener ae 28
INaturalidispersallicrte oA trceehiveyelanccss chsvanctaad Ges ented ote ators Gaaieiniensn eect aOR ie eee 29
Dspersalkby mane cess se soe) geyehacs se evans Sistekale tia ioe ve heteuel otenneeiin Sebieedy ost lee Ss TERRE Ceo nae 29
Natural hazards
Predatorsyes sscs.nsieacsus ene ene severe, sicte sete roverbetenste! ertvaiieeycrense: Seetents cae aE or EE Seno 31
Ectoparasitoids).% 2525 sf Seen iiersyn ce seta cok etna aa cc ae te ea ney ee en 32
Endoparasitoids eae or fh sis oc rats ate ies eine oafe ee Rie eet res aren oe eae RNR foe eo 34
Ep piParasitoidss = Sis cicve <cac ster, s eeec ate hare ea ta Sree eee ene Ae 34
Pathologicalihlazards™. 26 Siiices socio os esta cove ante ore ay ean eps ee nae pee 34

Piterature!Citedh esi ee Fed BAN eh | 2 a he ee ec tt teal ee nace icy non eat ee ee 35

Part 1

Origin, Relationships and Integumental
Anatomy of the Insect Order Embiidina

Preface

This is the first issue of EMBIA, a new publica-
tion which, at least initially, will serve as the outlet
for my large number of long-shelved manuscripts on
Embiidina. Many of these accumulated because I
had intended them to be chapters in a series of mono-
graphic volumes. However, at this late stage in my
life, this will not be possible.

Over the years, of necessity, laboratory research
was interrupted by expeditions—one lasting eighteen
months—to assemble a meaningful collection of the
order. Besides my curatorial duties at the California
Academy of Sciences, I was, of course, distracted by
many other activities, such as making available to
popular and educational media, the rich collection of
nature photographs I made during expeditions. In-
come from such activity paid much of the cost of the
fieldwork and, now, a portion required to produce
EMBIA.

Recent availability of desktop methods of pub-
lishing has encouraged me to discontinue scattering
my contributions in a diversity of periodicals. If I
live long enough (2000 is my 85th year!), many is-
sues of EMBIA—mostly on systematics—will ap-
pear in rapid succession.

I am using Embiidina (Embidina, Hagen, 1861),
the first group name applied exclusively to embiids,
as the order’s name in preference to more recent
Embiodea (Kusnezov, 1903) and Embioptera
(Shipley, 1904). Embiidina has had important early
use in monographs and is now rapidly supplanting
the awkward, inappropriate name Embioptera.

Acknowledgments

These will briefly serve this and future issues of
EMBIA. This anatomical treatment is in part based on
my unpublished Ph.D. dissertation (U.C. Berkeley,
1941). During this early work, I benefited from many
contacts with the late G. F. Ferris of Stanford Univer-
sity, as is evident in the style of my drawings. Follow-
ing four years of Army service in World War II, during

which I collected embiids in New Guinea and the Phil-
ippines, I began a series of extensive embiid surveys in
all major centers of the order’s evolution. The result-
ant collection, now numbering more than 300,000
specimens, representing about 1000 species (75%
undescribed) is almost overwhelming.

In this essential fieldwork I was helped by hun-
dreds of individuals, too numerous to mention. Ma-
jor financial support, in addition to my own funds,
came from the John Simon Guggenheim Founda-
tion (Amazon Basin), the National Geographic So-
ciety (Africa and Indo-Australia), the National
Science Foundation (field and laboratory activities),
the Dietz Fund, San Francisco Foundation (Africa),
the Schlinger Foundation and Health and Habitat.

Special thanks should be given to my field assis-
tants, who, like myself, participated without salary.
They were: Wilda S. Ross, Sandra M. Ross, Evert I.
Schlinger (second South American trip), Robin
Leech (first African trip), David Cavagnaro and
Clark Ross (Indo-Australia), Kenneth Lorenzen (sec-
ond African trip), Alan Stephen (third African trip)
and Mike Irwin (fourth African trip).

Later Robin Leech collected cultures for me es-
pecially in southeastern Asia and David Cavagnaro
did the same in Central America and Ecuador.

At the California Academy of Sciences, hundreds
of embiid cultures were maintained by the late Peter
Rubtzoff. Currently, with a Schlinger Foundation
grant and my personal funds, the Entomology
Department’s very able secretary, Julieta Parinas, as
off-duty employment, is typing manuscripts for
EMBIA and preparing pages for publication. Ear-
lier, also with Schlinger funds, Andrei Sourakov, as-
sisted me in many ways. I am also greatly indebted
to Jarmila Kukalova-Peck of Carleton University,
Canada for valuable information on wing
articulation,venation and Paleozoic relationships of
embiids. Norman D. Penny, and Katie Martin, Cali-
fornia Academy of Sciences, kindly checked manu-
scripts and made editorial improvements.

Ne

OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FiGurE A. Except for size and coloration, all female embiids have similar appearance. This photograph clearly

shows the spinning foretarsi and the enlarged hind femora within which large tibial flexor muscles motivate quick
reverse movement. New genus and species, Oligotomidae. Thailand. Body length 12.0 mm.

FiGuRE B. Typical adult male. Like females, alate males are similar in appearance throughout the order.
Antipaluria caribbeana Ross, Clothodidae. Venezuela. Body length 13.0 mm. Frontispiece shows spread
wings.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 3

Origin and Relationships

Summary

No proven pre-Tertiary embiid fossils have been
found but, according to Jarmila Kukalova-Peck (pers.
com.), there is evidence that Embiidina and what she
regards as its sister order, Plecoptera, evolved from a
common early Carboniferous ancestor. Tertiary em-
biid fossils closely resemble extant species—all with
the same adaptations easing backward, predator-es-
cape movement in narrow, silk galleries.

This rather “textbookish” treatment of integumen-
tal anatomy emphasizes these order-defining struc-
tures. Especially important are the peculiar wings of
males. To attain ultimate freedom of movement in
galleries, females of all species, by the process of
neoteny, became apterous (Fig. A).

Much attention is also given to evolution and di-
versity of the abdominal terminalia of males which
are of prime importance in defining species and high-
er taxonomic categories. Such diversity, as well as
other anatomical features, are profusely illustrated.
Additional information relating to anatomy will ap-
pear in EMBIA, Part 2, “A review of the biology of
Embiidina.”

Relationships

Based on a phylogenetic assessment of embiid
wing venation and articulation, using specimens pro-
vided by me, Kukalova-Peck has concluded that em-
biids comprise a very old order derived from
generalized plecopteroid stock and that this stock,
represented by Upper Carboniferous fossils, must
have served as a stem group for Plecoptera and Em-
biidina. Incidentally, she regards plecopteroids as
the sister group of all other Neoptera.

Earlier, Hennig (1981) speculated that ancestors
of Embiidina diverged from the paurometabolan line
about the beginning of the Carboniferous period. He
believed that the earliest division of Paurometabola
was Embuidina and Orthopteromorpha with the lat-
ter dividing into Blattopteriformia and Orthopteroi-
dea. Thus, he concluded that embiids have an isolated
systematic position and high categorical rank.

Rahle (1970) decided that there is a sister-group
relationship between Embiidina and Phasmida. Mat-
suda (1970) believed that Embiidina and Phasmida
are apomorphic sister groups of Plecoptera (Fig. C).

Kristensen (1975), while reviewing the phylogeny of
all hexapod orders and devoting considerable atten-
tion to Timema (Fig. D), a generalized phasmid, con-
cluded that determination of the relationship of
Embiidina is inconclusive. He believed that further
comparative study of the thorax of alate male embiids
(Fig. B) and their genitalia might aid determination of
the relationship of these orders. Storozhenko (1997)
placed Embiidina in the superorder Plecopterodea with
orders Plecoptera, Grylloblattida (Fig. E), Dermaptera,
and the extinct order Protelytroptera. Phasmids were
placed with Orthoptera in the superorder Orthopterodea.
The most thorough treatment of orthopteroid relation-
ships is that of Flook and Rowell (1998). They decid-
ed that Embiidina is a sister taxon to a clade consisting
of Phasmida and Orthoptera.
Origin

Embiid ancestors may have had the following
characteristics: (1) A prognathus head with chewing
mouthparts and long, filiform antennae, as in extant
embiids. (2) Legs moderately long, as in Plecoptera;
tarsi five-segmented. (3) Wings present in both sex-
es, subequal in size and shape except for a relatively
large anojugal area in hind wings; veins uniformly
sclerotized, none developed as blood sinuses. (4) Male
abdominal terminalia symmetrical, cerci more than two
segmented. (5) Habitat and food (as today), weath-
ered outer bark and dead leaf litter. (6) Defensive
biting weak; repugnant secretions absent. (7) Distri-
bution, equatorial portions of Laurasia, perhaps dur-
ing the Carboniferous period.

The ancestors probably ventured out of crevices
and other protective places to feed and retreated into
them to avoid predators and other adversities. It is
likely that short hind leg adaptations for rapid back-
ward movement developed during the order’s early
evolution. Such predator avoidance would have been
greatly improved if feeding forays could have oc-
curred beneath a silken cover or within galleries.
Such an unlikely adaptation was yet to evolve.

During the Carboniferous period there must have
existed embiid stem-group species with the above-
described characteristics and habits but, by chance,
only in one of them did mutations appear which en-
abled individuals to spin silken coverways using high-
ly efficient “tools’—the basitarsi of their forelegs.

4 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

One may well question, as with many other “thresh-
old innovations,” how the initial mutation could have
been effective. It may be that a single mutation might
have simultaneously modified each of the many
secretory pores on the plantar surface of the basal
segment of the foretarsi. If this had happened, the
quantity of silk immediately might have been suffi-
cient to produce at least crude predator-barriers ex-
tending outward from crevice retreats, or for crevice
coverage during vulnerable periods such as ecdysis,
or brood development.

Offspring of the mutant individuals could have
survived to reproduce in greater numbers, especial-
ly, as is likely, the brood had remained together and
mated incestuously within their parent’s “nest.” Over
time, through natural selection, the secretory cells
became multiple, elaborate internal glands while the
derm evaginated to produce associated tubular, seta-
like ejectors (Figs. 19, 20). This now-universal en-
semble has enabled all instars of all species of the
order to extend their feeding range within tubular gal-
leries, often enhanced in bark inhabiting species by
an additional sheet-like covering.

The tubular form of the galleries, sized to fit the
maker, is responsible for specializations which de-
fine the order, as follows: (1) A linear, supple body
with short legs. (2) Prognathism. (3) Greatly en-
larged hind tibial flexor muscles motivating rapid
backward movement in galleries. (4) Reduction of
wing-snag by flexibility resulting from desclerotiza-
tion of most wing veins. (5) Compensating ability
to stiffen wings for flight by blood pressure in sinus
veins—especially in the anterior radius (RBS). (6)
Universal apterism of females, and males of some
species, through neoteny—the ultimate adaptation for
improving backward movement within galleries.

Fossil record

The few fossils which unquestionably represent
Embiidina are modern forms from Tertiary forma-
tions and, therefore, useless for tracing the origin and
early evolution of the order. Fortunately, however,
much can be learned from extant, plesiomorphic spe-
cies such as Amazonia’s Clothoda nobilis (Gerstaek-
er) and C. longicauda Ross. Also, even without
fossils, one can postulate a great age for the order
because it must have fully evolved its order-defining
features, its major higher taxa, and wide distribution
on Pangaea long before its fragmentation. The or-
der’s specializations are much too complex to have
evolved de novo on each separated tectonic plate.

FiGuRE C. Paired stoneflies (Plecoptera), Acroneuria
californica (Banks) from northern California. Although
there are vast anatomical and biological differences between
modern Plecoptera and Embiidina, it is believed that the
two orders diverged from the same stock during the Car-
boniferous period. The copulatory positions are very simi-
lar but in Plecoptera there are no adaptations for reverse or
rapid movement.

FiGureE D. Adult male of Timema californica Scudder
(Phasmida), from central California, on foliage of live oak.
Its Paleozoic ancestors have been considered possible rela-
tives of those of Embiidina.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1

FiGURE E. Adult female of Grylloblatta campodeiformis
occidentalis Silvestri (Grylloblattida) from Mt. Baker,
Washington (1280 m). Kukalova-Peck (pers. com.) on the
basis of many characters concluded that grylloblattids be-
long to the blattoid line.

Pretertiary fossils

Zeuner (1936), treating his Germanoprisca zim-
mermanni, a new species from a European Permian
formation, concluded that Embiidina were derived
from the family Lemmatophoridae of the Protoper-
laria. These fossils belong to the plecopteroid stem
group which possess similar, generalized wing ve-
nation and attachment. No pretertiary fossil shows
blood sinus wing veins which characterize embiids
but, of course, much of the order’s evolution must
predate the appearance of sinus veins.

In 1937 Tillyard established the new suborder
Protembiaria (he used Embiaria as the ordinal name)
and family Protembiidae to include his Protembia
permiana, the type specimen of which was found in
Elmo deposits, Lower Permian of Kansas. Shortly
thereafter, Zalessky (1937) named two additional
species in his new genus, Tillyardembia, from relat-
ed Russian Permian deposits. Later, Carpenter
(1950), after re-examination of Kansas specimens,
decided that they do not represent Embiidina, and
assigned Protembia to the cumulative order Protor-
thoptera. His figure of P. permiana indicates that it
was a rather long-legged insect without any of the
specializations characterizing Embiidina.

In 1958 Martynova proposed a new embiid sub-
order, Sheimiodea, family Sheimiidae and species,
Sheimia sojanensis (Figs. 1, 2), based on a winged
fossil from the Upper Permian of Russia. Through
the kindness of Dr. Frank M. Carpenter I was able to
closely examine this specimen which, with two ap-
parently conspecific individuals, is preserved in a

Nn

piece of very hard, light tan rock. Anatomical de-
tails are not as clear as represented in Martynova’s
drawings (Fig. 2) and I can see no evidence that these
fossils represent Embiidina. The head and antennae
of Sheimia are embiid-like but the caudal cranial
margin appears to be more transverse, instead of
rounded, as inembiids. The apparently quadrate pro-
thorax of Sheimia is embiid-like but this character
isn’t peculiar to embiids. The pterothorax is well
sclerotized and broad, not unlike that of embiids, but
the scuta, figured as triangular by Martynova, are not
clearly so in the type specimen. The meso- and met-
athorax, however, seem to be nearly equal in size
and this embiid feature is associated with wide-
spaced, nearly equal (homonomous) wings.

Se
F Gi

Oa Vt ae Ve Be

FIGURE 1. Holotype of Sheimia sojanensis Martynova,
Upper Permian of Russia.

FIGURE 2. Martynova’s representation of the holotype of
Sheimia sojanensis (Redrawn by Ross).

6 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

Martynova’s drawing (Fig. 2) indicates that the
femora of the foreleg are greatly enlarged (an embiid
feature), but it is not certain that the faint impres-
sions she regarded as femora are indeed femora. The
wings are quite embiid-like but broader than typical.
The venation, although delineated as definite in
Martynova’s drawing, is obscure, yet there is a slight
indication in the specimen of the radius-border lines
which are peculiar to embiids. If indeed these were
present, this would be a strong basis for assigning
Sheimia to Embiidina, or a related extinct order.

The fossil’s abdomen is broad, perhaps due to
pressure of fossilization, and appears to be weakly
sclerotized with a darker apex. Unlike embiids, this
apex tapers to an acute point and Martinova’s draw-
ing indicates an ovipositor-like internal (?) structure,
but this could simply be a disconnected basal por-
tion of a filament-like cercus. She also indicated what
appears to be fragments of multisegmented cerci in
the vicinity of the abdomen’s apex.

After serious consideration I have concluded that
Sheimia is not an embiid. Carpenter (in litt, 1965)
drew the same conclusion and later (1976) placed
Sheimiidae and four other “problem” families in
“Families of uncertain ordinal position.” However,
there is a strong possibility that it belongs to the
embiid stem group.

Tertiary fossils

Perhaps the oldest Tertiary embiid fossil, Burmitem-
bia venosa Cockerell (1919) is a very small male (body
length 4.4 mm) preserved in amber (Burmite), pos-
sessed by the Paleontology Department, British Natu-
ral History Museum. Some workers, e.g., D. A.
Grimaldi (pers. com.), consider Burmite to be clearly
Eocene but others, e.g., A. P. Rasnitsyn (pers. com.),
believe it to be late Cretaceous. The fossil, definitely
an embiid, was treated in detail by Davis (1939). I
briefly examined the specimen, didn’t have time to
study it in detail, but I have decided to assign it to a
new family. Contrary to expectations, the species is
highly apomorphic.

Several embiid specimens have been encountered
in Baltic Amber (Eocene?) and, fortunately, two of
these are adult males clearly exhibiting the terminalia.
Thanks to the Geologischen Staatsinstitut in Ham-
burg, I have studied these specimens and suspect that
they are the cotypes of Embia antiqua Pictet (1854).
I assigned the species (Fig. 36) to a new genus,
Electroembia Ross (1956), and concluded that it is

more closely related to Rhagadochir Enderlein and
related genera of equatorial Africa than to Embia.
Apterism and melanism of these males suggests that
the climate during the period of entrapment in amber
was characterized by a long, dry season. In the Med-
iterranean region there are several species of Embia
and Haploembia with black, apterous males with gen-
eral appearance almost identical to E. antiqua and
can be distinguished only by examination of the ab-
dominal terminalia and hind tarsi.

Two small, alate embiids, collected in Chiapas
Amber, Mexico, are new species of the large modern
genus Oligembia Davis, family Teratembiidae, of
which there are many species in the American trop-
ics. Chiapas Amber has been “positively dated as
Oligocene and Miocene with possibility in some ar-
eas of ‘Eocene epoch’ ” (Hurd, Smith and Usinger,
1958).

More recently embiids have been collected in Oli-
gocene amber mines of northern Hispaniola (Domin-
ican Republic). Most of these fossils, some in my
collection, are winged males of the genus Oligembia.
One of these in the American Museum’s collection
became the holotype of Oligembia vetusta Szumik
(1994). Other Dominican Amber embiids belong to
two new genera in Anisembiidae, one of which is
represented today in Hispaniola, southern Mexico and
Central America.

Embiids preserved in East African copal, such as
Oligotoma westwoodi Hagen, represent modern spe-
cies of the family Teratembiidae.

The Eocene-Oligocene alate male fossil, Embia

florissantensis Cockerell, found in Florissant (Colo-

rado) volcanic shales, a large alate male of the fami-
ly Embiidae, has been assigned to a new genus,
Lithembia Ross (1984). The all-important charac-
ters of the male terminalia are indistinguishable.
According to A. P. Rasnitsyn (pers. com.), of
Moscow, the Miocene shale fossil Clothonopsis mi-
ocenica Hong and Wang (1993) isn’t an embiid but
has proved to be a bibionid fly, perhaps genus Plecia.

Integumental Anatomy

“Anatomy is what you see with your
eyes, morphology is what you think you
see with your mind.”

R. E. Snodgrass

It follows that the appropriate title of this
work should be the anatomy and
morphology of the Embiidina.

Introduction

For this simple exposition of the basic integu-
mental anatomy of Embiidina, a representative spe-
cies of the order, an Oligotoma, was selected because
three species of this genus have been widely distrib-
uted in commerce and, because adult males are at-
tracted to light, they are most frequently collected
and available for comparative study.

Adult females throughout the order are neotenic
and thus, except for their paragenital sternites, inter-
nal genitalia, size and coloration, are similar to early
stage nymphs. Probably due to advantageous endo-
crinal controls curtailing development, most features
of adult females are those of an early instar nymph
before even a trace of developing wings appear (Fig.
A). As will be discussed later, intermediate stages of
neoteny have evolved in adult males of many spe-
cies, as is shown by all ontogenetic stages of wing
development from complete aptery (Figs. 34, 35, 36)
through various stages of wing pad growth, brachyp-
tery and microptery. Fully formed wings, however,
are the most apparent characteristic of adult males of
the majority of species (Figs. B, 53). The alate con-
dition is associated with a full compliment of adult
embuid anatomical features, many of which certain-
ly must also have been possessed by adult females
during early evolution of the order.

Because of the high degree of specialization for
silk gallery life exhibited by adult females, it is safe
to assume that such features evolved not especially
to foster adult male survival but instead to secure and
prolong the lives of the more essential adult females.
Being short-lived and needing only to quickly dis-
perse and inseminate females, males faced less se-
lective pressure toward perfect adaptation to gallery

life and consequently are most likely to retain struc-
tures, such as wings, which are no longer of critical
importance in survival.

Among other plesiomorphic characters, adult
males also possess structures which aid location of
mates and insure successful copulation. Because of
their more complex anatomy, and consequent great-
er value in systematics, I will devote most attention
to the anatomy of adult males.

Methods

To fully expose sclerites which are partially or
entirely concealed in integumental folds, almost all
studied specimens were macerated in heated 10%
potassium hydroxide (KOH) and cleared of body con-
tents. Whenever possible drawings were made of
specimens positioned in alcohol. At times acid fuch-
sin stain was used to intensify details. Distortion of
position of sclerites is inevitable in KOH-treated spec-
imens and this must be considered during interpreta-
tion of certain drawings. Unless otherwise credited,
all of these were made by me, based on camera luci-
da and photographic outlines. I adopted the style of
R. E. Snodgrass (1935) with membranous areas stip-
pled and degree of sclerotization represented by thick-
ness of lines and some shading. Usually setae are
omitted. The figures are intended to be self-explana-
tory and the text only a guide and a means of calling
attention to the functional or phylogenetic significance
of structures. Unless otherwise indicated, they are
based on specimens of Oligotoma nigra Hagen.

Knowledge of musculature is vital to any study
of integumental anatomy, but I haven't had the time
nor background to make such a study. Therefore, I
asked Edward L Smith to at least investigate the mus-
culature of the abdominal terminalia of adult males

8 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

and some females. Some of this research has been
applied by me in the present paper. Perhaps his com-
plete research will be published separately.

Iam indebted to Christine A. Nalepa, a cockroach
specialist, for useful suggestions concerning my treat-
ment of embiid wings, their reduction and loss.

While concurrently producing Part 2 of EMBIA
which reviews the biology of Embiidina, I became
increasingly aware of the difficulty of separating be-
havior from anatomy—especially in the case of em-
buds where there is such a close relationship between
structure and function, as is most apparent in my treat-
ment of embiid wings. Consequently, some infor-
mation in this paper will be repeated in my review of
embuid biology.

Head

The head of embiids is orthopteroid in type, basi-
cally similar in all stages of all species (Fig. 3). Due

occiput

~~~~ sulcus

B male ventral

|} _--mandible ___

“>}--- anteclypeus - -

anterior
---~-tentorial pit --\-—-

Xo > epistomal sulcus GA

dorsal

posterior
--—tentorial pit ~

occipital foramen

postoccipital

to the universal vegetative diet of nymphs and adult
females, there has been no great need for head spe-
cialization and diversification. Only in adult males
are there significant cranial and mouthpart special-
izations and these relate to the habit of males of many,
if not all species, to grasp the female’s head with the
mandibles prior to, or during, copulation. Except in
highly neotenic males, guts of adult males are emp-
ty and thus their mandibles aren’t used for eating,
instead being used only as mating claspers and for
cutting an entrance into a gallery occupied by a re-
ceptive female.

Non-neotenic males also have varied cranial and
eye structure as well as longer more sensory anten-
nae (Figs. 4, 34). Such characters probably are relat-
ed to a male’s more frequent movement outside of
galleries to disperse and locate a mate. Due to neo-
teny, adult females retain the nymphoid head anato-
my of early instars of both sexes (Fig. 3C, D) but, as

D female ventral

FIGURE 3. General structure of head of A, B adult male and C, D female of Oligotoma nigra Hagen.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1

FIGURE 4. Diversity of heads of adult males (labrum not figured in some species). A. Embia. B. New genus
Embiidae. C. Enveja. D and F new genera, Anisembiidae. E. Pelorembia tumidiceps Ross, Anisembiidae. G.
and H new genera in a new family. I. Oligembia n. sp., Teratembiidae. J. Australembia, Australembiidae. K.
Metoligotoma, Australembiidae. L. New genus and species, Oligotomidae.

10 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

will be later discussed, males of some species also
have similar nymphoid heads due to neoteny. How-
ever, at least in one exceptional case, a species of
Metoligotoma Davis, adult males with completely
nymphoid heads appear as intraspecific variants in
populations which include males having an adult-type
head, normal for the species.

The head of Oligotoma nigra Hagen, as in all oth-
er species of the order, is strongly prognathous, an-
gled downward at about 35° (Fig. 13). Obviously,
prognathism, an adaptation easing movement in nar-
row galleries, causes eyes, antennae and mouthparts
to have a functional forward position. There is also
obsolescence of most cranial sutures.

A feature apparently associated with prognathism
is the universally-present, broad sclerotic bridge be-
tween the posterior tentorial pits and the occipital
foramen (Fig. 3B-D). In many species this bridge,
here termed “‘ventral bridge,” is so strongly devel-
oped that all traces of its origin are lacking. In males
of Oligotoma, however, and more so in those of cer-
tain other genera, there is evidence that much of this
bridge is due to postoccipital sclerotization. The pos-
sibility that the medial portion represents a true gula
(a posterior extension of the submentum) was con-
sidered. However, in females and nymphs the sub-
mentum is clearly delimited basally by a transverse
suture between the posterior tentorial pits and thus
the bridge is not continuous with the submentum. A
sclerotic bridge behind the posterior tentorial pits fre-
quently occurs in other prognathous insects but in
such cases a true gula is present. To my knowledge,
the occurrence of a non-gular bridge in this position
does not occur in any other order.

Sutures related to original head segmentation are
practically obliterated on the venter (posterior por-
tion) of the head. The post-occipital sulcus can be
traced, however, by lines on either side of the occip-
ital foramen and these diminish anteriorly. In males
there is a slight indication that they parallel the ex-
tremities of the mesal extension of the posterior cra-
nial walls on the ventral bridge. They might therefore
be postoccipital.

The occipital foramen (Fig. 3B-D) is triangular
and its margins, although thickened, are not especially
modified for contact with the first cervical sclerites
midway on each side. The lateral margins are more
thickened posteriorly and gradually form a small in-
ternal ridge extending around the posterior end of the
head to the dorsal surface where it diminishes. This

line coincides with the outer margin of the long later-
al maculation of the pattern visible in the young and
adults of certain species.

In adult males, the dorsal (frontal) surface of the
cranium usually lacks sutures (Fig. 3A), but in fe-
males, lines of weakness (“coronal sutures’’) serving
ecdysis, appear as very faint unpigmented lines ex-
tending forward to the middle of the dorsum of the
head where they fork as two broadly divergent, equal-
ly-faint lines (Fig. 3C). These terminate just before
reaching the bases of the antennae.

Both dorsally and ventrally, the cranium often has
a characteristic pattern which appears to be quite
uniform throughout the order and is visible in
nymphs, and adults of some species, as lighter pig-
mented areas. Melanization of most adults generally
completely obscures these paler areas, but close ex-
amination reveals that they can still be traced as ar-
eas lacking setae and reticulate in sculpture. This
pattern correlates with interior attachment of bundles
of mandibular muscles. Another frequent pattern is
a diffused, often golden, transverse “cloud” from eye
to eye on an otherwise dark head. This overlays the
brain and may have some special function, one per-
haps related to light perception, or to mating.

The anterior tentorial pits open dorsally as trans-
verse, short slits situated in the epistomal sulcus just
behind the anterior articulation of the mandible. The
epistomal sulcus is represented laterally only, but
serves to distinguish the clypeus from the frons. The
anteclypeus (Fig. 3A, C) is entirely membranous and
limited anteriorly by a transverse fold.

The antennal socket is surrounded by a basal
flange delimited by an incomplete antennal suture.
The space between the eye and the mouth cavity is
greatly reduced and its sutures are obsolete.

The tentorium of the male is here figured from
the anterior and posterior aspects (Fig. 5). The ante-
rior and posterior arms unite medially and form a
thin, quadrate corporotentorium. The anterior arms
bear small dorsal branches which appear to be vesti-
gial. The posterior arms are dilated and join the very
large crassa, or hypostoma, and the inwardly-slant-
ing lateral flanges of the submentum. In males these
lateral flanges are heavily sclerotized, whereas in
females they are submembranous. The general struc-
ture of the tentorium appears to be constant through-
out the order. The posterior tentorial pits are situated
on either side of the base of the submentum.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 1]

anterior articulation
of mandible

anterior tentorial
pit.

vertex

antennal socket

eye -----2‘

posterior articulatior

crassa - of mandible

submentum

lateral flange —

A cephalic aspect

anterior tentorial
arm ~_

_-eye

posterior tentorial

~ corporotentorium

cranial wall --~~

B caudal aspect

FIGURE 5. Tentorial structure of adult male of Oligo-
toma nigra Hagen.

Eyes

Paired compound eyes are the only organs of sight
or light perception, ocelli being absent in all species
of the order. The eyes of adult males of Oligotoma
nigra (Fig. 6), as in those of many other non-neoten-
ic species, are relatively large, inflated, reniform and
composed of numerous, large, convex facets. The
greater development of eyes in males usually corre-
lates with the presence of wings and thus activity
outside of galleries. Some alate males, however, have
small eyes. Very large, inflated eyes with prominent
facets, occur in males of many pale species which
disperse nocturnally. This combination of pale or
somber body coloration and large eyes is, of course,
a condition found in many nocturnal insects, as well
as in nocturnal vertebrates.

_ In females and nymphs throughout the order the
eyes are very small and have relatively few facets.
Such a condition in adult females is nymphoid and is
associated with almost complete confinement in silk
galleries where activities are probably guided more
by touch than sight. Males of well-pigmented spe-
cies, especially those which disperse diurnally, may
have smaller eyes often with dark pigment in the facet
interstices. In a few species the head is nearly hol-
optic; the space between the eyes being very narrow
and the post-ocular cranial bulk greatly reduced (Fig.
4B).

FIGURE 6. Scanning Electron Micrograph photo of eye of
adult male of Oligotoma nigra.

Antennae

The antennae of both sexes are basically similar
throughout the order. They are annulated and fili-
form, thus similar to those of Grylloblatta and nu-
merous other orthopteroid insects.

An antenna consists of a scape, a pedicel and
many flagellar segments (flagellomeres). Few spec-
imens can be found with a complete number because
terminal segments frequently are broken or bitten off
by other embiids. However, these may be partially
regenerated if lost during an early nymphal instar.
In general, the number of flagellar segments seems
to correlate with body size; the largest number, about
32, is present in adults of large species and the small-
est number, as few as 11, is found in tiny species of
the family Teratembiidae. Adults of average size,
such as those of Oligotoma spp., tend to have about
19 flagellar segments. Antennae of the first instar of
all examined species are nine-segmented, while those
of the second are twelve-segmented. This increase
is accomplished by the basal flagellar segment di-
viding into three.

Antennae of adult females are nymphoid and sim-
ilar throughout the order. However, non-nymphoid

12 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

adult males show consistent intergeneric and inter-
specific variation useful in classification. Such char-
acters involve segment proportions, type of vesture
and color. In many genera, and in some congeneric
species, the apical segments may be abruptly and con-
sistently white in both males and females. In many
genera, e.g., Archembia Ross, females have white-
tipped antennae while males of most species of this
genus have uniformly brown segments. Great an-
tennal length in large males is due in part to increased
segment-number, but also to segment-elongation.

As is to be expected in insects that venture into
open environments to locate a mate, antennae of
males have the greatest sensory function. Antennae
of males of two species were studied by Slifer and
Sekhon (1973) with specimens provided by me. They
reported at least five types of sense organs on flagel-
lar segments of Ptilocerembia sp. which, incidental-
ly, has the most hirsute antennae found in the order.

They concluded that the large sense organs are
“thick-walled chemoreceptors” (or contact chemore-
ceptors). Each has about five neurones at the base
and extend their dendrites to the tip of the hair with-
in a long cuticular sheath, at which point they are
exposed to air.

Mouthparts

The mouthparts are typically orthopteroid and
need not be described in detail. The labrum (Fig. 7)
is a simple, semicircular flap, slightly apically
emarginated, and membranous medially along its an-
terior margin, and clothed with setae. The ventral
surface, or epipharynx, is entirely membranous and
characterized by two widely spaced, nearly parallel
rows of inwardly-directed short setae. The surface
between these rows has a fine reticulated structure
which continues on the dorsal lining of the esopha-
gus. Well-developed tormae are located near each
basal angle of the labrum, and extend well within the
anteclypeus. The hypopharynx (Fig. 8) is large, sim-
ilar to that found in other orthopteroids. Its dorsal
(anterior) surface lies against the epipharynx and is
clothed with dense, scale-like setae.

Mandibles of nymphs and adult females (Fig. D-
F) are stout, asymmetrical, and have large,
multidentate grinding surfaces well adapted for chew-
ing coarse food. The points of dorsal (anterior) and
ventral (posterior) articulation are located medially
in nymphs and females, whereas in adult males they
are located near the caudal angle of each mandible.

-.- palatal - surface -.-

FIGURE 7. Labrum of female Oligotoma nigra. A. dorsal
B. ventral

The mandibles of adult males (Fig. 9A, C) gener-
ally greatly differ from those of females, as well as
those of their nymphs the mandibles of which are
identical to those of adult females. These differenc-
es are quite constant within a taxonomic group and
thus important in systematics. Unlike those of fe-
males, mandibles of males are often elongate, rather
flattened and without grinding surfaces. Their teeth
are few in number and located apically, usually with
three on the left mandible and two on the right. At
times these apical, or incisor teeth, may be fused and/
or curled ventrad (as in the genus Embia Latr.). In
all Anisembiidae, and certain other taxa, there is fu-
sion of the subapical teeth with the apicals.

In males of many species the inner face of each
mandible has a prominent, acute point sub-basally.
This, which may be the proxadental cusp, separates
the incisor area from the molar area, which is often
greatly projected mesad. Between the incisor teeth

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 13

esophagus

lacinia

FiGureE 8. Lateral aspect of hypopharynx.

and the proxadental cusp the surface may be smoothly,
inwardly arcuate but often it is produced as an
obtusely angulated point, here termed medial flange.
The outer, sub-basal angle of the mandible is
occasionally extensively produced. The diverse forms
of male mandibles are illustrated in Figure 4, as well
as in my publications dealing with embiid systematics.

Mandibles of adult males usually differ so great-
ly from a type adapted for a herbivorous diet that
some workers erroneously concluded that they must
be used for predation. Actually, the peculiar mandi-
bles of many species function as claspers to control
the head of females during copulation. Only in a
few species with highly neotenic males are the man-
dibles ever used for mastication of food because
males normally stop eating during the penultimate
instar. It is possible that the mandibles of males may
be effective in warding off small predators and rival
males. They are also useful for snipping an entrance
into silk galleries likely to contain a prospective mate.

Females and nymphs regularly use their mandi-
bles to pick up and place fecal pellets outside of gal-
lery walls. They are also used by ovipositing females
of some species to pulverize feces and habitat mate-
rials for inclusion in a paste deposited around eggs.
Such pulverizations may also be deposited on the
outer surface of a silk labyrinth to enhance or per-
haps obscure the protective cover. The mandibles
may also be used to gnaw burrows into the habitat—
soil, bark, dead twigs, etc.

The maxilla (Fig. 10) have no peculiar features
and are similar in both sexes throughout the order.
The maxillary palpi of adult males of Australembia
Ross, however, are exceptionally large and may as-
sist mandibular clasping during copulation. The base
of the cardo articulates well within the head.

The labium (Fig. 11), although generalized, ex-
hibits interesting modifications. The glossae once
were believed to be supplementary spinning organs
(Enderlein, 1912). This is not the case, however, and
no trace of labial gland openings have been found
associated with them. The mentum is greatly reduced
and is almost obsolete in adult males of many spe-
cies. In neotenic, apterous adult males and females,
as well as in nymphs, it is a small triangular sclerite.

The submentum is a well-developed broad plate,
weakly sclerotized and uniform in females and young,
but in adult males of Oligotoma Hagen (Fig. 11A)
and those of many other genera, it is heavily sclero-
tized with the anterior and lateral margins folded
inward, or inflexed with lateral flanges and fused ba-
sally to the posterior tentorial arms as described be-
fore. In some genera, such as Antipaluria Enderlein,
the submentum of the male is deeply bifoveate and
these depressions may be a source of glandular se-
cretions associated with mating. In Archembia Ross
the surface may be flat and from which tiny pores
produce a secretion that becomes a white coagulant
around setae in alcohol-preserved specimens. Males
of some species of a new southeast Asian genus have
a peculiar, rugose, conate submentum but in other
species of the genus the structure is normal. The pos-
terior tentorial pits are located at the proximal angles
of the submentum and are separated from the occip-
ital foramen by the extensive ventral bridge.

Cervix

The large size of the cervical sclerites (Fig. 12 B,
C) appears to be related to vigorous head activity.
For example, strong musculature is involved in head-
pushing to shape silk galleries as they are spun and
adult males of many species need cervical strength
to pull a female’s head to the right in a mandibular
grip prior to copulation. Adult males are also able to
turn their heads at surprisingly extreme angles as, in
a mantid-like manner, they follow the movements of
an observer, most often by males resting on walls
under a light.

14 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

">>, incisor teeth
4

y dentes
¢

---medial flange

. 5 Cee
incisor teeth
dentes

proxadental - - -
cusp

za Se ree ——_,’ ~~ extenso
con
condyle Syaducton muscle ae cs ue 1 tendon
(pre artis) aiachment pre artis) | attachment
(rectotendon ) Post artis
ks incisor teeth — - --
~~ incisor dentes ‘\
FY teeth y
ae proxadental
~ flange SEP
. _ Proxadental zy
cusp Zi
molar
Ke area
~molar \
‘ area ‘
/ \
/
- dorsal
\
\, /
adductor /
muscle iv
y, attachment / b
/ adductor ' exten
condyle / muscle attachment ! tendon’
(pre artis) post artis ~~~ condyle (pre artis) attachment

FIGURE 9. Mandibles of adult male (A—C) and adult female (D-F).

The first lateral cervical sclerites contact subme-
dial points on the sides of the occipital foramen but
these points are not especially modified to receive
such contact. The outer edge of each cervical scler-
ite is thickened and its anterior apex bears a small
hair plate. In the head’s normal position, about 35°
downward from the horizontal axis (Fig. 13), the sec-
ond lateral cervical sclerites are almost perpendicu-
lar to the longitudinal axis of the first. This angle
becomes obtuse as the head projects forward and is
thus more horizontal. Each second lateral cervical
sclerite articulates with a small, medio-ventral pro-
jection of the adjacent pronotal flange.

In the ventral neck membrane, immediately be-
fore the prosternum, are located small sclerites; the
anterior of which is smaller than the posterior (Fig.
12B). Crampton (1926) designated the first as an
intersternite and the second as the presternum (an an-
terior detachment of the prosternum). In the present
work these are simply designated anterior and poste-
rior ventral cervical sclerites. Ahead of these, between
apices of the first cervical sclerites, aré two membra-
nous domes which are densely setose in Clothoda
Enderlein. Similar setose domes are found in mem-
branes lateral to these apices. These are termed latero-
ventral cervical sclerites.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 15

lacinia
\

cranial
articulation

FIGURE 10. Maxilla, adult male.

paraglossae

NaN

DD NAW
erie ae

1) \4Z)4
RAN SLULY eae Iiy A
JL) ! TY)
Vy yh yy pd 1 y77, /

\ Vy lay;
YL G1,
NANA CHU LIL PAT oA
WAS OW GL h2

\/
A

|
cH hy

|submentum

FIGURE 11. Labium, adult male (A), adult female (B).

There is no great intra-ordinal variation in the cer-
vix except in males of certain genera. In Enveja
Navas the lateral cervical sclerites are especially
broadened and this seems to be associated with the

need for greater anchorage of large muscles moving
an exceptionally large head. In this case the second
lateral cervical sclerites are greatly produced me-
sad, almost touching the presternal sclerite.

Prothorax

The prothorax (Fig. 12A, C), one of the least
specialized parts of the body, is similar in both
sexes throughout the order; the prothorax of the
alate male simply being less robust. The generalized
condition reflects a lack of unusual prothoracic in-
volvement in embiid behavior.

The pronotum is unusual, however, in that it is
not produced laterally in folds and is usually much
narrower than the pterothorax. It is divided across
the anterior third by a deep furrow which delimits a
narrow anterior part, the prescutum, and a large con-
vex posterior part. The latter is divided medially by
a longitudinal suture probably associated with ecdys-
is. In males the posterior angles are broadly pro-
duced caudad as gradually diminishing extensions.
Medially, the caudal margin is developed as a small
point. In many cases, especially in nymphs and fe-
males, the pronotum bears a characteristic pigment-
ed design which lacks setae and is reticulated, as in
the head pattern. This design, like that on the crani-
um’s vertex, probably is related to internal muscle
attachment. In males of many species the prothorax
often is pale in coloration, or reddish in some diur-
nally dispersing species.

The very clearly-defined straight sutures which
form the lateral margins of the pronotum delimit flap-
like flanges on either side. These invaginate poster1-
orly to form two wide-mouthed, pleural apophyseal
pits, the apophyses of which are stout, conical and
directed caudad at a 45° angle. Submedially, a small
ventral lobe is produced against which the base of
the second cervical sclerite articulates. Behind this
projection a suture extends to the mouth of the apo-
physeal pit and defines the catepisternite as a dis-
tinct sclerite somewhat as in Plecoptera. This sclerite
is slightly convex and produced ventrad in the poste-
rior half to form the dorsal coxal articulation. The
trochantin is dorsally separated from the catepister-
nite by a transverse suture.

The prothoracic sternum (Fig. 12B) is a broad,
quadrate plate, the posterior angles of which are pro-
duced as short truncate processes. Narrow, lateral
areas separated by sutures may be subcoxal elements
fused to the prosternum. Slit-like openings of the

pronotal
flange
y x
pleural
cpophyseal
pit
X

prescutum

Shou SS
episternite,
pronotum
scutum

acrotergite

A

pronotal flange _

catepisternite _ ~

, oa / trochantin

a a
lateral *
cervical sclerites

posterior ventral
cervical sclerite

OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO.

pronotum

'
'
'
abe

. 149

latero- ventral
cervical sclerite~

hair plate
y

caterer - latero- dorsal
ventral cervical sclerite
sclerite — feesaal
--- cervical
posterior _ sclerites
ventral —- 5 a - Pa
cervical
sclerite

trochantin
\

prosternum

sternal

sternal

poststernum apophysis

B

pleural apophysis
1

, acrotergite
ii

spiracle

'
'
1
'
}

>

1 7
sternal apophysis
'
/
U /

prosternum poststernum

C

FiGureE 12. Structure of prothorax, adult male. A. Dorsal aspect, B. Ventral aspect, C. Lateral aspect. Oligotoma nigra

Hagen.

sternum’s apophyseal pits are located on either side
of the posterior margin. These are transverse and
develop internally as short, terminally-dilated, flat
apodemes. The broad area between the inner ends of
the apophyseal slits represents the point of basal con-

tact of the very large first poststernum which is fused
to the prosternum.

A detailed treatment of the prothorax, including

its musculature, will be found in Bitsch and Raymond
(1970).

Pterothorax of males

The pterothorax of alate males (Figs. 13-16) ex-
hibits a number of very interesting features. Unlike

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 17

FIGURE 13. Lateral aspect of head and thorax of a male showing angle of head in repose. Somewhat distended as a result

of KOH maceration. Oligotoma nigra Hagen.

corresponding somites of females which are round-
ed and distended by a large, food-filled crop, these
somites are dorso-ventrally depressed and transverse-
ly rectangulate in cross-section in alate males.

Males of Oligotoma nigra have a rather apomor-
phic type of embiid pterothorax but adequately repre-
sent general features. Most prominent is the large,
elongate, triangular mesonotum (scutum 2) which
abruptly arches downward anteriorly forming a near-
ly vertical prescutum terminating in a short phragma.
The acrotergite (anterior notal plate) is quite similar
to that of females and nymphs, being separated from
the antecostal suture by a membrane. The anterior
notal wing processes are prominent, strongly devel-
oped, with deep submembranous emarginations be-
hind them. The posterior notal wing processes, located
not far behind the former, are nearly obsolete, being
represented only by the anterior angles of very weak-
ly-sclerotized lateral flanges of the mesonotum. The
mesonotum is rather strongly arched in cross-section
and the lateral flanges are limited mesad by a low point
of the sides of this arched portion and hence are di-
rected upward. In apomorphic families, such as Olig-
otomidae, Anisembiidae and Teratembiidae, these
lateral areas are very weak. In plesiomorphic groups,
such as Clothodidae and Embiidae, they are more
heavily sclerotized. When wings are in repose, a cer-
tain amount of inward bending occurs along this area
and the tendency for its weakness may be an adapta-
tion for life in narrow galleries, for such bending per-
mits the wings to rest more directly over the body (Fig.
B). Otherwise, wing-edge projection beyond the
body’s lateral lines could cause disadvantageous fric-
tion with adjacent gallery walls. The most peculiar
feature of the pterothorax is the very great reduction
of the scutellum which causes the axillary cords to

almost meet on the mid-dorsal line. To my knowl-
edge, these are the longest axillary cords to be found
among insects.

The acrotergite (anterior notal plate) of the me-
sothorax is exceptionally large and closely contacts
the scutellum and the metathoracic prescutum. It is
divided medially by a longitudinal membranous cleft.
Small, elongate sclerites located on either side be-
neath the axillary cord probably represent isolated
rudiments of the anterior angles of the acrotergite.
Essentially, the metanotum has the same structure as
the mesonotum. It differs chiefly in its shorter pro-
portions and the fact that its acrotergite forms a bridge
between the scutellum and the first abdominal ter-
gum to which it is fused.

In males the pterothoracic pleura (Fig. 16B) are,
of course, more extensively developed than in females.
A peculiar feature is the nearly vertical prealar suture
near the anterior end of the episternum which delim-
its a small sclerotic area (prealar sclerite). This area
may represent a fusion of one of the anterior precox-
al sclerites, which in females is separated from the
episternum. Another possibility is that it results from
secondary folding. The suture is represented inter-
nally as a strong ridge directed caudad. The dorsal
end of the prealar suture forms a process separated
from the pleural wing process by a membranous area.
The basalare bridges the gap between these two pro-
cesses in the form of a small, elongate, convex arch
which is firmly attached at either extremity. The sub-
alar, a small unmodified sclerite, appears to be simi-
lar in males throughout the order.

The sternum of the pterothorax (Fig. 15B) is a
composite plate resulting from fusion of precoxal
sclerites (free in females and young) with the basis-
ternum and furcasternum. This fusion can readily be

18 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

scutum

2

mesonotum

CX

scutum

metanotum

abdominal

]

| as |

A

f

--spiracular lobe-- — -

,
pteralia =

B V lateral flange__
axillary cord --
--—-- episternum - -—-

----pleural suture - -
——---— epimeron - - -

i scutellum ~ - -

ig :|_ ~ 7 ~ acrotergite-- =
=4------ prescutum - -—

Nee pteralia- ~~ FA
" a ~~

—— -episternum - - -

~-pleural suture--
—-- epimeron - --

acrotergite- —

~~= laterotergite

abdominal ~
spiracle ]

scutum

2

mesonotum

scutum

3

metanotum

abdominal
tergum

abdominal
tergum

2

B

coxa’3

FiGuRE 14. Meso- and metathorax, dorsal aspect of A. Adult female, and B. Adult male. Oligotoma nigra Hagen.

traced by lines of internal thickening, by the position
of the sternal apophyseal pits, and often by distribu-
tion of setae. The sternum of a male of Oligotoma
nigra has been figured alongside that of the female
(Fig. ISA, B) and a comparison clearly indicates the
homologies of the various areas of the composite ster-
num. The subcoxal area, very prominent in the me-
sosternum, is obsolete in the metasternum of this
species, but is more evident in more plesiomorphic
genera, such as Embia.

The pterothorax of more plesiomorphic Embiid-
ina, such as clothodids and embiids, seems to retain

many primitive features, e.g., broader and shorter pro-
portions of the scutum with stronger lateral flanges,
a broader scutellum, a partially separated prescutum,
undivided acrotergites (notal plates), as well as cer-
tain features of the sternum.

Meso- and metathorax of females

The meso- and metathoracic anatomy of adult fe-
males of Oligotoma nigra (Figs. 14A, 15A, 16A) typ-
ifies that of adult females and nymphs throughout the
order, as well as that of fully neotenic (apterous) males.
Adult females unquestionably once possessed wings.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1

\

Poststernum

spina

basisternum

furcasternum

spina

basisternum

Their present simple juvenile thoracic structure strongly
contrasts with ancestral complexity retained in the tho-
rax of alate males. Thoracic simplicity in adult females
obviously results from neoteny for its sclerotization is
nearly identical to that of immature stages of females,
and the second instar of males destined to develop
wings. Some difficulties in interpretation can perhaps
be attributed to the fact that most sclerites are rather
soft and are seldom fused along sutural lines. This
would appear to make possible distention of the midg-
ut and more supple body movement—the latter a de-
cided advantage in gallery life. The apterous condition

apophyseal pit 1

--- anapleural cleft ---

—— --— preepisternum - —

- preepisternum —

|_ - - -ventropleurite ~ -|,.
- apophyseal pit 3 - --

---------coxa 3

~~ abdominal sternum 1~

prealar suture -- —

\
ny

spina

basisternum

furca-
sternum

spina
basisternum

furcasternum _

episternum -

trochantin- - - - -

B

FicurE 15. Meso- and metathorax, ventral aspect of A. Adult female and B. Adult male. Oligotoma nigra Hagen.

of adult females, and the same tendency in males of
some species, is undoubtedly an adaptation for life in
silk galleries, as will be fully treated in the section deal-
ing with wings.

The meso- and metathoracic scuta of females are
simple, large plates, each having a prominent anteri-
or acrotergite. The narrow transverse sclerite in the
membrane between the anterior margin of each and
its acrotergite is probably homologous to the fully
developed prescutum in alate males. There is no great
development of phragmata. The posterior margin of
the metathoracic scutum closely contacts the first

20 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

prescutum scutum 2 acrotergite prescutum scutum 3
H

. \ .
acrotergite pleural suture epimeron 2
. \ ‘

1
i)
1 f
'
'

spiracle
\

iy 1 abdominal abdominal
y \ pleural suture | — epimeron3 terga spiracles
iy \ \ o “4 .

Hat , | h 1 \ v uy

1

prealar ~~ 3
sclerite ~ if Ci ay ! \ \ , ha ogy \
U CRUE SL ! ' spiracle —/ ee \ \
‘ \ .
Poststernum 7 Lo ' coxa 2 ! Bonet Z / —\\ \ trochantin 3 S \
; / asi sternum Fs .
episternum = / J ! } P 4 \\ apophyseal pit \ abdominal
- * _ \
preepisternum |” apophyseal pit ' preep ister nun ‘ episternum \ Sternum 2
basisternum poststernum furcasternum coxa 3
Prescutum pleural suture .
; . Prescutum pleural suture acrotergite abdominal
acrotergite lireare : “ ; : tergum spiracles
'g | Wing base acrotergite epimeron a is
\

w
a
ie
=
i
|

; i
‘ I
prealar suture | '
1 !
' ‘

1
1 1
\ n '
\ \ !
1
1 ' ,
1
\

1 1
spiracle | | reepist fia Spi i <
1 | Preepisternum | coxa2 _ spiracle : preepisternum \ coxa 3 M
1 I Y
/ 1 ' ! i ‘ :
poststernum | basisternum i ; Nees abdominal
|! trochantin basisternum trochantin sternum

prealar sclerite

B

FIGURE 16. Meso- and metathorax, lateral aspect of A. Adult female and B. Adult male.

abdominal tergum without any apparent development
of an acrotergite of the latter.

The pleura of both thoracic somites are similar (Fig.
16A). The pleural suture dorsally contacts the anterior
angle of the scutum and extends ventro-caudad at an
angle of about 30° to form the dorsal articulation point
of each coxa. Coxae are located at points directly in
line with the posterior angles of the scutum. This dor-
so-ventral compression, characteristic of all embiids,
probably is associated with life in narrow galleries for
it results in a lengthening of the body and a decrease in
its vertical plane. On either side of each pleural suture

is a narrow epimeron and a slightly broader epister-
num. The mesothoracic episternum produces a ven-
tral trochantin which, in the metathorax, is delimited
by a deep cleft extending nearly to the pleural suture.
The apices of the tronchantins constitute the ventral
articulation of the coxae.

Between the ventral (anterior) margin of the epis-
ternum and the basisternum there are large, weakly
sclerotized plates which I believe are pleural in origin.
In the mesothorax they are rather rhomboid in shape,
unequal in size and, as in the metathorax, are two in
number, probably as a result of secondary division. In

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 21

the metathorax they are unequal; the anterior sclerite
(pre-episternum) being larger and greatly elongated,
the posterior (ventropleurite) smaller. In an alate
male’s thorax, the ventropleurite is fused to the sides
of the furcasternum and the pre-episternum becomes
the anterior appendix (prealar area) of the adult male
episternum. Beneath each coxa there is a small quad-
rate sclerite surrounded by membrane which, in adult
males, also becomes a part of the sternal plate. Ten-
tatively, I regard this series of three sclerites as the
precoxal arc.

The sternal region (Fig. 15A, B) is one of the most
interesting portions of the thorax of Embiidina. Mat-
suda’s (1960) interpretation of the thoracic sternites
of insects seems to satisfactorily explain the condi-
tions present. The development of the sternum in
the order is one of the best examples of the broad
type characteristic of orthopteroid orders. This is
manifested by widely-spaced sternal apophyseal pits
and a consequently large furcasternum between them,
as is best exhibited in the mesothorax.

In the mesothorax of females the sternal apophy-
seal pits are located in the lateral membranes of the
sternum and are widely separated by the broad fur-
casternum which, in combination with a greatly de-
veloped basisternum, covers most of the venter of
the mesothorax. The sternal apophyses are very slen-
der and project only a short distance inward.

The sternum of the metathorax is similar to that
of the preceding somite but shorter, with its apophy-
seal pits very close together. The pits may have “mi-
grated” caudad from lateral positions comparable to
those of the mesosternum. The sternal apophyses
are much stouter and longer than those of the me-
sothorax and are strongly directed dorso-laterad.

The first pair of thoracic spiracles is located in
the intersomital membrane behind the prothorax on
prominent, setiferous, lateral lobes and they are ac-
companied by small sclerites. The second pair of
spiracles is subventrally located near the anterior an-
gles of the metathoracic basisternum. They are in-
conspicuous and do not open on prominent lobes.

Barlet (1985) has published a more detailed treat-
ment of the thorax.

Legs

The legs of embiids, remarkably similar in all
taxa, are very short relative to body size as best ex-
hibited by nymphs, adult females, and neotenic adult

males, which are more perfectly adapted to gallery
life than alate, slightly longer-legged, non-neotenic
adult males. Also universal is the distinct form and
function of each pair of legs. Most significant are
the unique foretarsi which produce the peculiar en-
vironment and thus the order’s associated anatomi-
cal and behavioral adaptations. It is noteworthy that
all leg features, even the spinning glands, are similar
in all instars of all species of the order.

Prothoracic legs

Silk-spinning is the most important function of
the forelegs and, accordingly, all of its segments are
enlarged, well sclerotized and muscled to serve vig-
orous activity (Fig. 17). The resting position of the
forelegs is forward, tarsi paralleling the sides of the
head. During spinning the legs sweep very wide
arcs—even back and up to the mid-line of the tho-
rax. This activity includes short strokes in many
directions combined with outward pushes of the tar-
si and head to shape galleries. Even teneral individ-
uals spin soon after ecdysis. Diminutive forelegs
regenerated after loss during a nymphal instar are
also capable of spinning.

The spinning organs (Figs. 18, 19) have been de-
scribed several times (e.g., Melander, 1902; Rimsky-
Korsakov, 1914; Mukerji, 1928; best by Barth, 1954;
and Alberti and Storck, 1976). The contention of
Enderlein (1909, 1912), who probably never observed
a live embiid, that the silk is produced by labial
glands, has long been rejected.

The glands are confined to the greatly enlarged
basal segment of foretarsus and number at least 150
per tarsus. Variations according to taxa and instar
remain to be studied. Each gland consists of a large
lumen enclosed in an irregular layer of syncytial cells.
The globular, but often quadrate and irregular shaped
glands, are closely appressed to each other. They are
sufficiently large to be visible under low magnifica-
tion, especially through the pale thin derm of the plan-
tar surface. The crowded glands resemble seeds in a
pomegranate fruit.

Within each lumen there is a chambered corbicu-
lum (ampulla) with radiating filaments which appar-
ently direct liquid silk secretions of gland cells into a
duct (one per gland). Each duct opens distally at the
tip of a relatively long, hollow filament, here termed
silk-ejector (ejector of Barth, 1954). They are not
homologous to setae (Fig. 20A—C).

in)
Nm

tarsus ~__

trochanter ~~

C tarsus of mesothoracic leg

FiGuRE 17. A. Prothoracic leg, B. Mesothoracic leg, and
C. Its tarsus.

N N77 -----. ejectors

VENTRAL

FiGuRE 18. External aspects of foretarsus of a female.

OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

According to Barth’s excellent cytological inves-
tigation—the first made with the aid of an electron
microscope—the silk-ejectors and their glands are
unique organs. As shown in my diagrammed hypoth-
esis (Fig. 19A), the glands perhaps evolved from in-
vaginations of secretory cell pores in the ectodermis.
Each silk-ejector apparently represents a setae-like,
evagination of the exocuticle which might first have
been simply a cuticular rim around a pore opening.

It is probable that the early ectodermal glands
were hollow “balls” composed of walled secretory
cells. The duct, or constriction, leading to each pri-
mordial silk-ejector must at first have been fully cell-
lined. The inner walls of these duct cells appear to
have gradually thickened and fused to become the
elaborate ducts which now extend from the silk glands
to the ejectors. Similar ectodermal glands are treat-
ed and illustrated by Snodgrass (1935: 62, fig. 32).

Cells of the ectodermal gland and the duct later
became syncytial. Only a small amount of nucleated
cytoplasm persists on the outer wall of a duct. The
ducts are thread-like and wend their way between
the glands with never more than one duct per gland
and its ejector. The evolution of such unique, com-
plex organs is difficult to comprehend unless one con-
siders the probability that even small mutations must
have had simultaneous expression on all glands.
Under such circumstances, even minute, favorable
modifications could have had almost immediate sig-
nificance in the survival of the possessor.

The basal tarsal segment (Fig. 18) is elongate-
oval. In alate males it is narrower and more cylindri-
cal in cross-section than in females, nymphs and
nymphoid apterous males. In females and nymphs
the segment is almost triangular in cross-section due
to slant of the inner-dorsal surface. The outer side is
almost vertical. The inner-dorsal surface is almost
entirely submembranous, finely reticulate in texture
and almost lacks setae. A medial, longitudinal de-
pression which extends the length of the surface grad-
ually becomes stronger terminally. The area
immediately inward of this depression often pulsates.
Dorsal surfaces adjacent to this membrane are darker
and clothed with moderately long setae of the usual
type.

The ventral, or plantar, surface of the basal seg-
ment (Fig. 20A) is entirely membranous, pale, often
pinkish in color, and densely covered with short,

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 23

7 ~~~exocuticle
silk
~~ejector

seueace
| oaeeses: |
(eee: Socs ets

Ficure 19. A. Hypothetical origin of silk gland. B. Defini-
tive structure of a single gland and its ejector. C. Glandular
content of foretarsus of a female.

ejectors

acutely-tapered microspicules which may function
as combs. Silk-ejectors on the outer edge arise with-
in small, circular areas lacking such microspicules.
The ejectors are characterized by a lack of basal sock-
ets and are thin walled, fragile and whitish— condi-
tions apparently associated with their hollow nature.
They are variable in form and length but are straight
except for occasional abrupt curvature of their com-
plex apices (Fig. 20B, C).

Along the lateral margins of the plantar surface
the microspicules are much shorter and the bare spots
from which the ejectors arise are more conspicuous.
Ejectors arising from these lateral borders appear to
be much longer than elsewhere and probably are the
most important silk-strand ejectors.

The mid-segment of the foretarsus is a small tri-
angular pad with a membranous ventral surface
clothed with both microspicules and silk-ejectors; the
latter being denser but, of course, less numerous than
on the basal segment. These ejectors are served by
ducts arising in glands located within the basal seg-
ment. The distal segment of the foretarsus is smaller
than those on the mid- and hindlegs but otherwise
anatomically similar. During spinning it is elevated
to prevent its claws from hooking into the webs.

Nature of embiid silk

Using specimens provided by me, K. M. Rudall
of the Department of Biophysics, University of Leeds,
England, made brief studies of the silk during 1973—
78 and in letters to me, conveyed the following in-
formation.

Double X-ray diffraction pattern shows embiid
silk to be of the classic Group I fibroin silk charac-
teristic of Bombyx mori and Beria sp. (Cymbidae).
This silk group isn’t found in most other Lepidoptera,
therefore, its occurrence in an embiid is “most excit-
ing.” He regarded the dermal tarsal glands of embi-
ids as perhaps simpler secretory systems than the long
salivary glands of Lepidoptera larvae. The diffrac-
tion pattern has been interpreted (by universal agree-
ment) to be due to alternative residues along extended
polypeptide chains of glycene and alanine. Some of
the alanine positions are reflected by serine (serine
is nearly the same size as alanine). Rudall noted that
the main difference between Bombyx mori silk and
that of embiids is that the serine to alanine ratio is
reversed, there being a much greater content of serine.
The ratios in the embiid sample was 197 glycine to
40 alanine to 130 serine.

Mesothoracic legs

The mesothoracic legs (Fig. 17B, C) are the least
specialized of the three pairs. Their most notable
feature is a relatively great reduction in size and they
do not seem to have much locomotor importance in
or out of galleries. In their normal position they ex-
tend laterad and span a gallery’s interior. They are
capable of great upward movement and frequently
the tarsus of one of the pair is able to contact the
upper surface of a gallery while the body is in a nor-
mal position. It is possible that these legs aid rota-
tion of the body within a gallery.

The tarsi consist of elongated, unmodified seg-
ments (Fig. 17C). The basal segment is evenly clothed
ventrally with stout setae. The mid-tarsal segment is
small and has a distal papilla. In adult females and
nymphs of many species, the ventral membrane of
this papilla is clothed with very small, basally-directed
peg-setae which must aid reverse traction on the silk
substrate. In adult males this surface never is echinu-
late and perhaps this is another indication of the male’s
poorer adaptation to gallery life.

24 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FiGure 20. A. Silk-ejectors (white) and “combs” of female Embia (SEM 440x). B. and C. Silk-ejectors and “combs of
female Embia (SEM 1750x). Ejectors are highly variable in length and distal structure.

Metathoracic legs

Formerly, there was some question as to the func-
tion of the greatly enlarged femora of the hind legs
(Fig. 21). It was assumed that their enlargement
indicated a saltatorial function. Davis (1936) exam-
ined the tibial muscles and noted that, unlike saltato-
rial insects which have large hind tibial extensor or
levator muscles, the flexor or depressor muscle is
greatly enlarged and thus accounts for the large size
of the femora. The extensor, or levator, muscle is
much reduced and fits into a groove atop the flexor
(Fig. 21C, D).

Such tibial musculature obviously facilitates
backward movement in the galleries—an activity re-
quiring strong muscles to flex the hind legs. Rapid
reverse movement in narrow galleries has been a ma-

jor factor in the evolution of many adaptive charac-
teristics of the order, notably wing modifications and,
ultimately, wing elimination in all females as well as
in males of many species. This will be more fully
discussed in the section dealing with wings.

Reverse movement can be rapid, very smooth,
the body axis remaining straight. In contrast, an em-
biid’s forward movement usually is rather awkward
and slow but, with stimulation, it can be very fleet
especially within the galleries. The resting position
of the hind legs is straight back, closely paralleling
the sides of the abdomen. In walking, the terminal
segment of the hind tarsus is elevated and doesn’t

contact the web and, obviously, this avoids a snag-
ging of the claws in the silk. Thus, substrate contact
of the tarsus is with the basal and middle segments.

The basal segment, or basitarsus, probably a com-
posite of three fused ancestral segments, is slender
in adult males, stout in nymphs and adult females
(Fig. 22). In both sexes the plantar surface bears large,
irregular, peg-like setae which are more slender in
males. In some females the setae are directed basad
in the distal half of the segment. The upper and lat-
eral surfaces are clothed with long, slender setae
which, like many others on the legs, have the outer
curvature finely serrate. Scanning electronic micro-
scope images show each serration as the apex of one
of the fibers composing a seta.

The distal end of the plantar surface of the basi-
tarsus is always produced as a membranous “blad-
der,” or papilla. Many species, however— often all
species of a taxon— have a submedial, second papil-
la on the ventral surface (Fig. 22B). This may repre-
sent one of the papillae of the three segments which
fused to form the basitarsus. If so, its presence may
be plesiomorphic. Two papillae are possessed by spe-
cies of Clothoda, the order’s most plesiomorphic ge-
nus. However, a second papilla may appear
sporadically throughout the order as a specific or even
generic character without any phylogenetic signifi-
cance. In spite of this, the presence or absence of a
second papilla is constant in all instars of a species,
usually so within a genus and thus serves as an im-

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART | 25

depressor - “
muscle

~-basitarsus

D cross section
°
hind femur

FIGURE 21. A. Hind leg of male. B. Hind leg of female,
C and D. Tibial musculature.

portant character in systematics. In a few species,
males have only one papilla, females two. Occasion-
ally, the second papilla is much reduced, often sim-
ply a small, unprotruded, clear spot. Such papillae
appear to be homologous to arolia of the tarsi of grass-
hoppers and other insects and were so designated by
Imms (1913). Snodgrass (1935:198) called them tar-
sal pulvilli, or euplantulae.
Wings

Ancestral embiids must have possessed fairly in-
flexible wings similar in texture to those of most oth-
er alate insects. However, with increasing dependence
on quick reverse movement in silk galleries as a prin-
ciple means of evading predators, the apices of such
relatively stiff wings must have frequently snagged
against opposing gallery walls and slowed or prevent-
ed escape. To overcome such a problem, and to in-
crease suppleness during U-turns, embiids long ago
evolved extraordinary wing flexibility. As a result,
when in repose over the back, the wings of all mod-

ern embiids readily fold transversely and slide for-
ward toward the head (Fig. 23), thereby reducing
likelihood of a snag, or ““barb-effect.”

Although the wings usually fold upward and
cephalad across their midline, they can bend at al-
most any point and may even irregularly crumple.
Such flexibility appears to have been accomplished
through desclerotization of most of the longitudinal
veins behind the radial blood sinus (RBS), notably
the media (M) and cubitus (Cu). Perhaps reduction
of plication is also involved, as suggested by the fact
that vein cuticularization is almost entirely confined
to the dorsal membrane with only blood sinus veins
evident on both wing surfaces.

FIGURE 22. A. Hind tarsus of first instar and adult of male
Oligotoma nigra. B. First instar and adult hind tarsus of
female Haploembia solieri (both showing medial papilla.
lacking in O. nigra).

26 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FIGURE 23. Wing flip of a reverse-moving male. New genus and species, Oligotomidae. Thailand.

The great flexibility of such wings leads one to
question how they could serve as flight organs. The
answer seems to lie in the functioning of four blood
sinus veins, the most important of which follows the
course of the anterior radius (RBS = RA). Second-
ary blood sinuses include the subcostal (ScBS = Sc),
the cubital (CuBS), and the anal (ABS = A) (Fig.
24). These sinuses correspond in position to the as-
sociated veins except for the terminal halves of CuBS
and ABS which do not contain tracheae. Through-
out the order the sinus veins are cuticularized on both
the dorsal and ventral wing membranes.

In effect, each sinus vein is a dark, glossy, dor-
soventrally-flattened sac that is tapered and perhaps
closed distally. When wings are in repose, sinus sur-
faces are flat but during “excitement” preceding
flight, hemocoelic pressure must increase and, as the
wings extend laterally preparatory to flight,
hemolymph (“‘blood”’) flows into the sinuses causing
their surfaces to become convex. Microscopic ex-
amination of a turgid sinus shows that hemocytes
synchronously pulsate with the beat of the dorsal
blood vessel and do not perceptibly move distad.

Together, the turgid subcostal, cubital and anal
sinuses appear to function as tines of a fork that stiff-
en the wing’s base, while the radial blood sinus (RBS)

stiffens almost the entire length of the wing’s lead
edge. When an embiid alights, and the wings return
to repose over the body, it is probable that a crimp-
ing of the blood sinuses occurs in the axillary region.
Hemolymph then gradually oozes back into the body
cavity.

One could regard, especially the longitudinal wing
veins of all insects, as narrow sinuses into which co-
elomic blood pressure extends as a means of expand-
ing wings from pad to adult form. Following a teneral
period, during which these veins gradually cuticu-
larize, blood pressure into them must decrease, or
completely cease except for pressure in the axillary
area which remains to extend wings for flight (“take-
off’), or defensive, or sexual, display, among insects
which fold their wings over the body when they are
not in use.

In contrast, blood sinus veins of embiids enable
blood pressure to continue on and off throughout the
adult life of all species with alate males. Over time,
such veins broadened and sclerotized to become dis-
tally-closed sinuses characterizing wings of all spe-
cies of the order. Embiids apparently are the only
insects—indeed the only animals on Earth—capa-
ble of temporarily stiffening, otherwise flexible,
wings.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBHIDINA, PART 1 27

/
CuBS
FOREWING

HINDWING

CuP CuA 2

-RAor RBS
=e RP

FIGURE 24. Wing venation of a plesiomorphic embiid, Clothoda longicauda Ross (Clothodidae). ScBS, RBS, CuBS
and ABS are symbols for blood sinus veins. RML = radius marginal line. Note: MP vein is anomalous in hindwing.

The fact that the sinus veins are similarly-deve-
loped throughout the order indicates a great antiqui-
ty for the specialization, but one preceded by
evolution of the silk-spinning ability and increased
survival associated with the complete confinement
in silk galleries. Such unusual wing adaptations must
have been initially perfected in a single species for it
seems unlikely that they could have evolved with
identical complexity more than once. Furthermore,
primary selection pressure for wing flexibility was
most likely on adult females, not adult males, to adapt
them more perfectly to gallery life. The adult life of
a male is too short and its prime biological function—
simply a non-feeding “vehicle” for delivery of sperm
and genetic diversity—is performed so quickly that
there would not seem to be sufficient selective ad-
vantage for males to have been the primary target for
this remarkable wing specialization. Adult females,
however, must live long enough to mature eggs, ovi-
posit and guard eggs and early-instar nymphs. Thus,
a Species more significantly benefits from adapta-
tions which foster quick, predator-avoiding move-

ment of females in narrow silk galleries. The evolu-
tion of alternating wing flexibility and stiffening rep-
resents one of the order’s first “attempts” to ease
backward movement.

Venation

Strength (cuticularization) and completeness of
venation (plesiomorphic features), are greatest in spe-
cies with largest body size (also plesiomorphic).
Complex venation apparently is related to the obvi-
ous need for more extensive blood and air distribu-
tion in larger wings, especially in the pads of large
nymphs during development.

Wings of Clothoda longicauda Ross (Fig. 24), one
of the order’s most plesiomorphic species, may be
used for interpreting venation. Most past work on
Embiidina, including mine, used Comstock-Needham
nomenclature of veins, as summarized by Comstock
(1918). Based on tracheation, a modified nomencla-
ture is here adopted. I am aware that many entomol-
ogists since Comstock have rejected tracheation as a
basis for interpretation; however, in embiids there is

28 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

consistent correlation between tracheation and ve-
nation, both in pads and in fully developed wings. It
should be noted that tracheae aren’t easily seen, if at
all, in wings of dead, dry specimens but are most
apparent in freshly killed individuals, especially those
still teneral (Fig. 25A, B).

The costa (C) forms the anterior wing margin and
terminates at the apex of the anterior radius (RA).
The subcosta (Sc) is short, cuticularized, terminates
within the wing’s basal fourth, and probably func-
tions as a blood sinus, hence the symbol ScBS. The
anterior radius (RA), the wing’s most prominent vein
and, most significant blood sinus (RBS), stiffens the
wing’s lead edge. It originates strongly near the wing
base and parallels the costal margin almost to the
wing’s apex, at which point it tapers and usually
curves downward to join RP. Especially in apomor-
phic species of Anisembiidae, RA slants toward the
costal margin well before the wing’s distal curvature.

Throughout most of its length RBS is exception-
ally broad, sclerotized, glossy, darkly melanized and
bordered, except at basal fourth, by peculiar streaks
(radius margin lines, RML), having a fleshy, granu-
lar appearance, and usually are brick red in color.
Images taken by SEM show the surface of these lines
to be irregularly wrinkled, perhaps to accommodate
alternate turgescence and flattening of RBS. The ra-
dial blood sinus and its margins would be an inter-
esting subject for detailed study.

Near the base of RBS a short stem juts caudad
and then immediately extends distad (RP + MA) for
almost a third of the wing’s length, at which point it
appears to fork. However, the anterior branch of the
“fork” is simply a continuation of RP, which never is
forked in any species of the order. Thus, the basal
portion of this “vein” really represents a fusion of RP
and MA. Superficially, especially in long dead spec-
imens, this composite stem appears to be a single,
cuticularized vein; however, two parallel tracheae
within it have separate wing-base origins, a condi-
tion which prevails in wings and wing pads of all spe-
cies of the order (Fig. 26). The posterior branch of
the fork, once designated R4+s, is here regarded as a
continuation of the anterior branch of the media (MA).
This vein usually is forked, forming MA1+2 and
MA3+. especially in plesiomorphic taxa of the order.

The stem of the “media” is fused to the anterior
edge of the sclerotic base of the cubital blood sinus
(CuBS). Separate tracheae, within this stem angle
abruptly forward toward the stem of RP and then sep-

arate to form MA and MP. MP then extends to the
mid-margin of the wing and is very rarely forked.

The trachea of the cubitus (Cu) at first parallels
those of MA and MP within the extreme base of the
cubital blood sinus (CuBS), but then angles caudad
within the sinus. At the sinus’ mid-length the tra-
chea and its vein emerge from the sinus and parallels
MP to the wing margin. In plesiomorphic embiids,
or those with exceptionally large wings, the cubitus
may be multibranched. I tentatively regard the basal
branch of Cu as CuP.

Beyond this diversion of Cu, the cubital blood
sinus continues its broad, tapered, diagonal course
to the wing’s hind margin. No trachea follows the
distal half of the sinus. The hyaline stripe between
CuBS and the anal blood sinus (ABS) tends to crease
and this suggests that it may be equivalent to the cla-
val suture which delimits the anal fold in wings of
certain other insect orders. It tends to fold upward
and forward in embids.

As stated before, the anal (or vannal) area of
embiid wings is much reduced, but it has a dark cen-
tral line, an anal blood sinus (ABS), within which
one can see the unbranched trachea of the anal vein
(A). Kukalova-Peck (pers. com.) prefers to desig-
nate A as AA because, in many other insects, there is
a posterior branch of A, therefore an AP vein. There
is always a cross-vein between A and the base of the
cubital blood sinus. Kukalova-Peck regards this as
an “anal brace.”

The hindwing is similar to the forewing but al-
ways is shorter, broader and certain veins, such as
MP and Cu, may be less strongly represented. In
some species of Archembia Ross, the anal area 1s
slightly more expanded than in the forewing (Fig. 27).

Cross-veins may be highly variable in position
and number within a species and may even differ in
the left and right wings of a single individual. How-
ever, their general positions and number is rather con-
stant within a species, or even a genus. In Oligotoma,
for example, cross-veins seldom if ever are present
behind RP. There seems to be no regularity of cross-
vein position which would justify nomenclature for
cells they delimit.

The upper and lower wing surfaces are densely
clothed with small, short hairs commonly called
microtrichiae which, having no apparent basal sock-
ets, appear to arise as direct outgrowths of the wing
membrane. The entire outer margins of the wing and

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART | 29

HINDWING

A

FIGURE 25. A. Dark field illumination of wings of freshly killed “Embia” surcoufi Navas (Embiidae) showing tracheae
(not visible in “dead” wings). B. Tracheation of teneral Archembia batesi (McLachlan) (Embiidae). Photo also shows
enlarged anal area of the wing occurring in some species of Archembia Ross.

30 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

=eNG
CuA

FIGURE 26. Tracheation of forewing pad of penultimate
instar “Embia” surcoufi Navas.

courses of all veins, even those unsclerotized, bear
rather large setae (macrotrichiae) arising from prom-
inent sockets. They are particularly large along the
costal margin. Large setae are also present on the
ventral wing membrane but are fewer in number and
without definite arrangement.

The wings of Pararhagadochir Davis (Fig. 28)
of the Embiidae exemplify the more apomorphic (=
reduced) wing venation found in most Embiidina.
In such venation the apex of MAi+2 and MA3+4 and
all of MP are unsclerotized, each vein traceable only
by its row of macrotrichiae and pigment stripe.

The wings of Enveja bequaerti Navas (Fig. 29)
represent contrasting, perhaps plesiomorphic. vena-
tion in which all veins are heavily sclerotized. but
the cubitus isn’t forked.

In some species of the apomorphic family Ter-
atembiidae, as well as in other taxa, small body size
correlates with vein reduction, including all veins ex-
cept those functioning as blood sinuses (Fig. 31).

In a common type of reduction, MA always is
simple (Fig. 30). All species of unrelated families
Anisembiidae and Oligotomidae have such reduc-
tion, a reduction which sporadically occurs in sey-
eral other distinct evolutionary lines, such as within
Embiidae, Notoligotomidae and Teratembiidae.

The most apomorphic wing of the order is found
in a South African new species of Teratembiidae. Its
wings are very small, slender, with all veins except
the blood sinuses obsolete, and the wing margins have
especially long setae. Such thysanoptery parallels
the tendency of the smallest species of various insect
orders (e.g., Certain parasitic wasps, small Tri-
choptera, some microlepidoptera and ptiliid beetles)
to have slender, fringed wings.

Although there is much convergence in wing ve-
nation in embiids, venational characters have impor-

tant, supplemental value in the definition of species,
genera, and even families. It is doubtful, however, if
wing characters can ever be used as the primary ba-
sis of phylogenetic conclusions.

Perhaps, because of the unimportance of flight
in evasion of predators, embiid wings exhibit con-
siderable random, often anomalous, intraspecific vari-
ation. The most extreme, yet consistent, wing
anomaly occurs in an Amazonian new species of
Oligembia Davis, which has normal forewings but,
not even a trace of hindwings (Fig. 32), and the met-
athorax is reduced to the size of an abdominal seg-
ment. However, very closely related species from
the same region have normal hindwings and thus the
hindwing atrophy is of no significance in systemat-
ics. If, however, a comparable character appeared in
certain other insect orders, it might become the basis
for proposing a distinct higher taxon.

Embiid wing anomalies seem to illustrate a law
in biology which may be expressed, as follows: if an
anatomical feature is not vital to survival or repro-
duction, it may be subject to much anomalous varia-
tion within a taxon. Thus, in flight-dependent insects,
such as most Diptera and Hymenoptera, wing fea-
tures are relatively constant and are, therefore, de-
pendable characters in systematics. The converse
appears to be the case in embiids because their flight
has little or no adaptive value.

Wing pigmentation

Apparently pigmentation of embiid wings always
is confined to the upper membrane, the ventral being
completely hyaline except for dark “imprints” of the
blood sinuses. Alternating longitudinal dark stripes
and hyaline intervals, although faint in some species,
are characteristic features of the upper membrane of
all embiid wings except possibly those of Burmitem-
bia venosa Cockerell, an Eocene (?) amber fossil from
Burma.

The veins and/or their macrotrichiae are centered
in the dark stripes; the intensity, width and marginal
definition of which are consistent within a species.
In turn, such melanization correlates with the overall
pigmentation of the male. In arid regions many spe-
cies disperse nocturnally and generally are pale tan
with wings correspondingly pale with faint venal
stripes and broad hyaline intervals, the margins of
which are often indefinite and irregular.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 31

RBS

RP

FIGURE 27. Wings of Archembia n. sp. (Embiidae) showing narrow hyaline stripes and broad anal area (a plesiomorphic
condition).

FIGURE 29. Forewing of Enveja bequarti showing strong venation and white cross-veins. The wing’s costal and hind
margins are golden.

32 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

RES

-RP

FIGURE 30. Reduced vein-strength of the forewing and venation of a species of Chelicerca (Anisembiidae). The unforked
MA vein characterizes many species and entire families, e.g., Anisembiidae and Oligotomidae.

CuBS’
\
Cu;,
FIGURE 31. Reduced vein-strength of the forewing of Teratembia geniculata (Teratembiidae). Many African species of
this family have vein MA unforked.

FIGURE 32. Complete atrophy of metathorax and hindwing loss in a new Brazilian species of Oligembia (Teratembiidae).
Terminalia characters indicate, however, that this isn’t a very distinct species.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 33

There are exceptions to this, however, for melan-
ism of some species Occurring in certain seasonally-
arid regions (e.g., Embia Latreille spp., in the
Mediterranean region) is often associated with nup-
tial dispersal during cool weather following winter
and early spring rainfall. In such cases melanism
may foster rapid body warming in the sun, enabling
more rapid dispersal movement, thereby reducing
predation hazard.

Especially in the humid tropics, wings of darkly-
pigmented males (e.g., Ptilocerembia Friederichs
spp.) often have a beautiful, metallic blue or laven-
der luster. This is especially intense on the veins and
diminishes toward the hyaline stripes.

In some species wing pigmentation may contrib-
ute to a mimetic appearance. Blackness causes the
wings to resemble elytra of aposematic beetles, such
as Pyrochroidae, which, like their embiid mimics,
usually have a reddish prothorax and black elytra.
In another type of mimicry, extensive golden mar-
gins of the wings (e.g., Enveja Navas spp.) result in
a resemblance to chemically-repugnant lycid beetles
occurring in the same habitats.

Curiously, cross-veins often are conspicuously
white when crossing hyaline intervals and darkly pig-
mented while crossing the dark stripes. The result-
ant white, cross-slashing of embiid wings is
characteristic of richly-pigmented, diurnal males
occurring in humid environments.

As mentioned before, brick red or pink, “granu-
lar” lines (RML = radius marginal lines) bordering
the radius blood sinus characterize all embiid wings.
This subcutaneous granular pigmentation may also
be present in the costal margin and in veins RP and
MA, as in Pararhagadochir Davis (Fig. 28). Each
side of a radius margin line may be pale. Sometimes
the granular red lines extend into adjacent longitudi-
nal veins and cross-veins.

Reflecting on the signifance of the universal strip-
ing of embiid wings, I have concluded that wings of
ancestral species must have been uniformly dark and
that the hyaline intervals evolved in lines of weak-
ness which fostered longitudinal plication. Observed
obliquely, embiid wings display slight ridges corre-
sponding to the veins, and furrows correlated with
the hyaline intervals. Unlike most insect wings, those
of embiids do not have a pattern of positive and neg-
ative longitudinal veins; the ventral membrane lacks
cuticularized veins. However, at least in some gen-

era, e.g., Clothoda Enderlein, veins are pale in color
on the ventral membrane.

In some clothodids, such as Clothoda nobilis
Enderlein and Antipaluria marginata Ross, the wing’s
costal margin is white. In some species of Anisem-
biidae and Teratembuidae extreme apices of the wings
are white.

Wing expansion following ecdysis

During most of the penultimate instar, wing pads
of males are flat with venation identical to that of
adults (not zig-zagged, for example, as in some Ple-
coptera). Nearing ecdysis, the pads become thick-
ened, convex and opaque white. When the exuviae
is shed, the pads at first retain the fleshy shape and
jut out from the thorax at about a 30° angle. This
probably assists flow of blood into the pads.

Within ten minutes the projected pads begin to
flatten and expand from the costal to anal margins.
Then they gradually assume the normal, repose po-
sition over the dorsum of the thorax. Periodically,
the embiid wriggles and rotates its body. Concur-
rently, the abdominal terminalia are distended and
the cerci project laterad at 45°—perhaps due to an
increase in hemocoelic pressure throughout the body.

In about twenty minutes the basal half of each
wing has fully expanded and flattened, the distal half
remaining as narrow and as fleshy as at the time of
ecdysis. In about thirty additional minutes the entire
costal margin has expanded and only the distal ex-
tremity of the posterior margin remains fleshy. This
condition prevails for another thirty minutes after
which the wings are fully expanded but remain white
with veins paler than the intervening membranes.

About two hours after ecdysis the wings have at-
tained their definitive shape and thickness and their
pigment stripes and hyaline intervals are faintly ap-
parent. Seven hours after ecdysis the wings are gray
in tone and the body and leg pigmentation is well
developed. After about twenty-four hours wing pig-
mentation, or cuticularization, is completed and the
male has ingested his exuviae. The male tends to
remain in one position for at least another day fol-
lowing ecdysis.

Wing articulation
Tam indebted to Jarmila Kukalova-Peck who, dur-
ing a visit to my laboratory (1999), greatly improved
my treatment of wing articulation. Details, presented

34 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

in Figure 33A, will be elaborated and possibly cor-
rected, in one of her future publications.

Articulation of the hindwings is similar to that of
forewings. There is, however, a trend toward a slight-
ly weaker representation of sclerites which is un-
doubtedly correlated with the wing’s smaller size.

Flexion of wings

In repose the wings lie flat over the back (Fig. B)
much as in termites, zorapterans and stoneflies. The
anal area of the wing is nearly obsolete and is repre-
sented only by a small, basal, posterior corner which,
in a fully-flexed wing, folds inward beneath the wing
surface. The important fold occurs at the wing base.
I have observed the mechanism of this basal flexion
in both living and dead specimens of Oligotoma ni-
gra and noted movements of the sclerites, as follows:
the fulcrum, or pivot, of the flexion is anterior: being
at the point of articulation of the anterior part of the
first axillary sclerite with the base of the subcosta.
An imaginary line drawn from this point through the
basal articulation of the third axillary, and still an-
other from it through the point of contact of the apex
of the third axillary with the posterior angle of the
anal band, delimit a narrow triangular area. As the
wing returns from the completely extended position
to rest over the back, folds occur along these lines
and the area becomes completely inverted. During
this movement (or the reverse) only the three axil-
lary sclerites change position while the other parts of
the wing-base remain stationary. During flexion the
first axillary rotates against the anterior wing pro-
cess through a 90° arc. The second axillary becomes
upright and the third is entirely inverted. The anteri-
or membrane is stretched around the fulcral point and,
finally, at least half of it comes to le parallel with
the side of the body.

The flexion of the wing thus seems to correspond
to that of many other insects. One point requiring
further investigation, in the light of the order’s wing
peculiarities, is the possible control of blood circu-
lation in the large radial blood-sinus (RBS), and oth-
er blood-sinus veins by means of movements in the
wing base. This may occur, at least in the case of
RBS, by simple pressure of the anterior membrane
against the fulcral point. There is a small strength-
ened point in the membrane opposite this fulcral point
which may fit across the place of strongest contact.

Flight

Because predator-avoidance especially depends
on remaining within silk galleries, flight did not
evolve as an important means of defense or dispers-
al. Adult males are slow to take flight and do not
readily fly away from a disturbance. They are more
likely to run away. In preparation for flight, a male
rises high on his forelegs, at times lifting them off
of the substrate, the head may bob up and down, and
the antennae may vibrate and rapidly twirl. Flight
distance usually is short, perhaps seldom exceeding
a meter and, soon after alighting, there may be a rep-
etition of the pre-flight and flight behavior. At times
take-off follows a short run or a hop.

Flight is a swirling, aimless, fluttering with ap-
parently no more directional contro! than that of nup-
tials of most species of termites. However, males of
nocturnally-dispersing species fly toward artificial
lights. It may also be assumed that they can direct
their flight to a gallery containing a receptive fe-
male. Especially in flight, diurnally-dispersing males
of certain species can easily be mistaken for various
aposematic beetles, such as some lycids, and
pyrochroids, which also serve as models for mim-
icry of many other insects.

Reduction and elimination of
wings

Obviously, the ultimate adaptation for rapid re-
verse movement in galleries is complete wing elim-
ination, now universal in females, through neoteny.
This also has occurred independently in males on
almost every evolutionary line within the order. A
similar neotenic “solution” developed among work-
er and soldier termites. Alate termites, however, break
off no-longer-needed wings prior to copulation and
a return to gallery life.

Due to friction, it is probable that more winged
embiid males are caught by predators at the extremi-
ties of galleries than are apterous, or subapterous,
members of a colony. Thus, mutations resulting in
wing-loss, or reduction in their size, are likely to be
selected. Another, and perhaps more significant fac-
tor, is that alate individuals are more likely to make
hazardous flights out of protective-galleries and
thereby become exposed to predation and other haz-
ards, such as desiccation. The fact that aptery is not
yet universal in males may be explained, as follows:

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART |

WwW
n

basivenalia
tegula BR,BM,BCu

humeral

RP+MA
scutum
anterior
MP
Ist axillary
sclerite —~—
2nd _ axillary
ieee
sclerite CuA
CuP

posterior wing

process :
¥
g
“See AA
’~ axillary cord

FIGURE 33. A. Wing attachment of forewing of a plesiomorphic embiid, Archembia kotzbaueri (Navas), Embiidae.
Nomenclature based on studies of Dr. Kukalova-Peck (pers. com.). Explanation of symbols: Sc = subcosta; BSc,BR,
BM, BCu, BA = subcostal, radial, medial, cubital, and anal basivenales; medial plate including medial (FM) and cubital
(FCu) fulcalares, is almost completely reduced. FA = minute anal arm of third axillary sclerite. The symbol AA (anterior
anal vein) is used because there is a tiny anal vein just caudad of the base of AA (not expressed in this species). AABS =
anterior anal blood sinus; CuBS = cubital blood sinus; RBS = radial blood sinus. Note that in this drawing the axillaries
are somewhat spread apart and that the distance between CuBS and AABS is exaggerated.

Ist axillary sclerite

i
'
1
!
'

humeral plate .

tegula, “ait axillary sclerite

axillary
cord

_ subalare
a

4 Se
a

te _basivenalia
— 7% BR,BM, BCu humeral plate
Sore

Ly

scutum 4 ; ie basalare._ ( sy
wa, pest \\ (|
<2 y om egula, WSN,

7 ais AN

C oa
anterior ee

wing process ~ 3rd axillary sclerite

. epimeron
posterior

wing process -~

episternum _-- pleural fold

FIGURE 33. B. Dorsal aspect of wing attachment of Oligotoma nigra. C. Pleural aspect of wing attachment.

36 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

1. Selective pressures favoring wing-loss muta-
tions are too short in duration to be effective. Life of
adult males is undoubtedly short (alate males take
no food) and they probably mate before the disad-
vantage of wings is fully felt. In other words, imper-
fect adaptation of alate males to gallery life is of no
importance to a species once males have mated.

2. It is also possible that factors which might
result in brachypterism, or apterism, of males have
not yet appeared for test in all species.

3. Wing possession may be less disadvantageous
in favorable environments, such as tropical
rainforests. In such habitats apterism of males is very
rare, whereas it is very common in regions with a
long dry season. It is thus likely that alate males dis-
persing in arid regions not only face exposure to ad-
verse climatic conditions, but also are more visible
to predators seeking prey in exposed arid habitats.

4. Wide dispersal of advantageous genes is se-
lected through retention of wings.

It may be that universal wing-loss in females was
not entirely due to selection against wing-possession
per se, but instead for a need to also eliminate pro-
jecting ovipositor structures which might have had
an even greater barb-effect than wings in reverse
movement. One could say that females lost their
wings as part of a nymphalization “package” that
resulted in a loss of most protruding adult append-
ages by cessation of development at an early nymphal
instar, perhaps not later than the second, before even
buds of such appendages make their appearance. This
neotenization was probably effected by mutations
influencing the secretion of the juvenile hormone.

Such a major specialization of wing structure and
function must have occurred long after the order
evolved most of its other peculiar specializations be-
fore fragmentation of Pangaea. It would be difficult
to explain the present widespread distribution of the
order if females had been apterous and needed to
hazardously extend their range afoot outside of pro-
tective galleries. Both sexes must have been alate
during the major evolutionary and distributional his-
tory of the order and female apterism must have later
convergently occurred on all evolutionary lines. The
appearance of more aggressive predators, such as
ants, could have been a factor favoring apterism and
confinement to galleries.

Apterism in males (Figs. 34, 35, 36) also has
independently appeared many times within the order,

but in varying degrees. Females can deposit eggs in
galleries without specialized oviposition structures;
males, however, with the exception of those of a new
family from Afghanistan, must have well-developed
genitalia and mandibles (often used as head claspers)
to insure copulation. Thus, even in highly neotenic
males, such as those of the Australian family
Australembiidae, neoteny mostly retards wing
development while, apparently through “tissue
competence,” the genitalia and head (to a lesser
extent), become fully adult. In some species of
Australembiidae, however, the head of adult males
is variable, an occasional individual within a
population has a head indistinguishable from that of
anymph, or of an adult female, and continues to feed.

Males of many species gain advantages of
apterism through degrees of thoracic neoteny. As a
result, males of some species have only wing buds,
or pads, in various states between the extremes of

a
(Petenescceceseeeeae™

(4664433.

FIGURE 34. Adult neotenic, apterous male (left) and fe-
male of a new species of Neorhagadochir Ross (Embiidae)
from an arid region of Nicaragua. Unlike its blackish
congeners, this species is pale ferrugineous due to its sub-
terranean habits.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIDINA, PART 1 37

FIGURE 35. Neotenic, completely apterous adult male of
Haploembia solieri (Oligotomidae) (slide preparation, body
length 10.0 mm). Endemic to the Mediterranean region.
Note: tong-like mandibles for grasping head of females
during mating.

full wing pad development and no trace of pads what-
soever. All this is probably due to different levels of
secretion and timing of juvenile hormone. In some
species there may be percentages of apterous,
subapterous, micropterous and alate males within a
species’ population, or in those of certain geographic
populations of a single species.

In arid environments apterous males are more
likely to remain within the parent colony, mate with
a sister, and thereby inbreed the wingless trend or
condition. Because of their greater ease of move-
ment, apterous males should survive in greater num-
bers and eventually male apterism could become
universal within a given population. Conversely, in
damper, more benign environments, any trend toward
male apterism might be swamped out by random
matings of alate individuals which are more likely to
survive flights from colony to colony.

FIGURE 36. Neotenic apterous adult male of Electroembia
antiqua (Pictet) (Embiidae). Baltic Amber, Hamburg Geo-
logical Museum, body length about 10.0 mm. This fossil
demonstrates antiquity of neoteny in males and close re-
semblance to modern species.

Some families include genera and species which
have radiated into marginal environments and thus
have apterous, or subapterous males. Therefore,
apterism in males must be used with great caution as
a character in systematic studies for it is most often
environmentally related. It is interesting to note that
one of the oldest known fossil species, Electroembia
antiqua (Pictet) of Baltic Amber (Ross, 1956), is com-
pletely apterous (Fig. 36) and this suggests that the
ancient land surface which supported the “amber
forest” might have experienced a long dry season. It
is possible that 1t was once in what 1s now a Mediter-
ranean, seasonally-dry global position, a terrain which
has since drifted northward into a colder, wetter lati-
tude.

Ultimate in the trend toward almost complete
neotenization of males is in a peculiar, undescribed,
subterranean species occurring in the desert steppes
of western Afghanistan. In this species males not

38 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

only have lost all traces of wings but also have com-
pletely nymphoid bodies and abdominal terminalia
(except for tiny rudiments).

There is an interesting correlation between wing
size and body proportions. Relatively large wings
are characteristic of light-bodied, slender, small-
headed males of species occurring at higher eleva-
tions of damp, equatorial mountains. In contrast,
shorter, narrower wings are possessed by robust,
larger-headed males living at lower altitudes and in
semi-arid regions. Except for slight venational diver-
sity, often involving vein-desclerotization, the wings
of embiids are remarkably similar in all species.

Abdomen

The abdomen is slender, usually as long as head
and thorax combined. In nymphs, adult females, and
apterous adult males of some species, it 1s circular in
cross-section. However, in alate adult males of most
species it is dorso-ventrally flattened due to reduced
content, all food having been excreted during the
penultimate nymphal stage. Fat storage is limited,
and internal reproductive organs are much smaller
than those of females.

Ten abdominal somites are conspicuous in both
sexes but vestiges of the 11th and 12th persist. Basic
somatization is most apparent in nymphs and adult
females whereas that of adult males is confused by
complexity of external genitalia, especially in
apomorphic taxa, as illustrated (Figs. 44-53).

The first abdominal tergum of nymphs and adult
females, a simple plate without an acrotergite, closely
contacts the metathoracic scutum. In alate males an
extensive, medially-cleft acrotergite is fused to the
metathoracic scutum. Terga of somites two through
eight are similar to each other but the ninth is much
shorter, broader, extends ventrad down each side of
the abdomen and almost contacts the outer margins
of the ninth sternum, or hypandrium (H). In nymphs
and females the tenth tergum is large, triangular, and
its outer basal angles extend ventrad to the sides of
the ninth sternum (Figs. 37, 38).

Matsuda (1976) regarded the produced apex of
the tenth tergum as the supra-anal lobe fused to the
tenth tergum. I have concluded that only the small,
weakly sclerotized area just beneath the apex of the
tenth tergum is a vestige of the eleventh (labelled
epiproct in Figs. 37 and 38). In females of some
species this vestige is separated from the apex of the

tenth by a transversely wrinkled, non-setose, inter-
somital membrane and the vestige bears its own se-
tae. Edward L. Smith informs me that intertergal
muscles are attached to this sclerite. It, and the apex
of the tenth tergum, develop into significant terminalia
structures, e.g., the epiproct (EP) and the medial flap
(MF), prominent in adult males of many species.

Located just beneath the lateral margins of the
first eight abdominal terga are elongate laterotergites,
each of which has a spiracle in the anterior end. In
many species the laterotergites of these somites are
divided into two sections, the posterior of which usu-
ally is much smaller (Fig. 38). Spiracles and
laterotergites are absent on somites nine and ten, the
positions of the latter being filled by latero-ventro
extensions of the terga.

In many apomorphic embiids, such as species of
Oligotoma, the first sternum is small and triangular
but in plesiomorphic genera, such as Clothoda and
Embia, it is larger and more transverse. Sternites of
somites two through seven are nearly equal in size
and form, each being subquadrate with a narrower
base. In adult females sternites of somites eight and
nine, which are associated with the vulva, are vari-
ously modified and will be separately discussed.

On either side of sternites three through eight there
are narrow, elongate sclerites similar in shape to the
laterotergites. Apparently these are abdominal
pleurites. They are almost entirely absent on the first
two somites, being represented only by two setae
adjacent to the sternite of the second somite. Pleurites
are absent on somites nine and ten. Inevitably, there
are difficulties in interpreting terminal abdominal
somites. Using Snodgrass (1935) as an authority, I
have decided that the paraprocts are structures of
somite nine. Edward L. Smith (pers. com.) believes
that the paraprocts are hemisternites of somite ten.
Matsuda (1976), however, regarded them as struc-
tures of somite twelve, suggesting that somite eleven
is greatly reduced with cerci remaining as its only
elements and, therefore, that structures of somite
twelve immediately follow those of somite ten. I
am not prepared to question these conclusions and
will simply endeavor to correctly homologize
terminalia characters without being overly concerned
with their somital associations.

The cerci of nymphs and adult females are simi-
lar throughout the order and seemingly comprise only
two segments (I realize that these are properly termed
flagellomeres but, for simplicity, I use the term seg-

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 39

ment). Cercus-basipodites, or coxites, may be re-
garded as basal segments of three-segmented cerci.
Each basipodite forms an almost complete fleshy ring
around the base of a cercus. The cercus segments
usually are elongate, cylindrical and often unevenly
sclerotized. Except fora relatively large nerve, most
of the content of a segment appears to be fat. The
cercus muscles are attached to the basipodites which,
in the left cerci of adult males of many species, be-
come fused to the base of the basal segment. In
plesiomorphic species the derm of the cerci is evenly
sclerotized, but in some apomorphic taxa it may be
almost entirely membranous. As a species character
the distal segment may be contrastingly pale or white
due to the color of tissue within a transparent derm.

The cerci bear setae of two types. Most numer-
ous are ordinary, tapered setae of various sizes which
arise from simple, circular sockets. These occur on
all surfaces but may be especially dense on the inner
faces of the basal segments. In males such density
may augment copulatory grip. The second type of
seta is finer, less tapered, arises from a rosette-type
of socket, or pit, and is most common on the usually-
less-sclerotized outer side of the basal segments. In-
variably throughout the order, the distal segment has
only one such seta on its inner side in a species-char-
acteristic position. Such setae are present in many
other arthropods and are often called trichobothria.
Probably both types of setae are mechano-receptors
providing tactile guidance, especially during reverse
movement in the galleries.

External genitalia of females

Because of neoteny, the external genitalia of fe-
males are underdeveloped. However, some species
possess buds of gonopophyses which might attain
adult form if females completed development, as they
did during their pre-neotenic evolutionary period.

Reduction of genitalia is possible because of the
simplicity of oviposition. A female merely attaches
eggs to a surface within the galleries, or on a silk
substrate, therefore no special structures are required
to insert them into a substrate. Furthermore, vital
reverse movements to escape predators would be
slowed, or arrested, if ovipositing structures protruded
and snagged against silk gallery walls.

The terminal abdominal terga, paraprocts, and
cerci of adult females (Figs. 37, 38) are identical to
those of nymphal instars. The only external evidences

of maturity, besides the open vulva, are slight modi-
fications of the eighth and ninth sternites. In
Oligotoma (Fig. 33), and many other genera, there is
no trace of valvulae, but between the eighth and ninth
sternites there is a slight, transverse, translucent ridge,
or carina. This elevation is subject to much modifi-
cation within the order; for example, in many spe-
cies of Embia its caudal side has two deep fossae
with glossy, sclerotic surfaces. The ninth sternite
usually has a baso-medial notch which varies from a
membranous condition to the dark, glossy, concave
sclerite found in some species of Embia.

In other genera, notably Metoligotoma Davis, ru-
diments of valvulae are quite conspicuous. In Meto-
ligotoma (Fig. 38) the eighth sternite is small and
lies beneath two blunt, fleshy lobes, or pads, which
bear small, rudimentary sclerites. These appear to
be rudiments of the first valvulae. Overlapping the
anterior margin of the ninth sternite, a prominent, bi-
lobed, non-setose ridge (probably a homolog of the
membranous ridge described for Oligotoma) is
present which may be a specialized rudiment of the
second valvulae. The lobes themselves may be rudi-
ments of the second valvulae and the low connecting
ridge represents anterior intervalvula. The ninth ster-
nite is small and has a large, quadrate membranous
area in the basal half. The pouch-like development
of this area, described for Oligotoma, is well devel-
oped in Metoligotoma and 1s partially concealed by
a vestige of the base of the second valvulae. The
aperture of the accessory gland may be located in
this pouch (Snodgrass, 1935, fig. 314B).

Although species of Clothoda have the most ple-
siomorphic males of the order, adult females of the
genus lack even traces of valvulae, the ninth ster-
num being simple and lacking a basal pouch.
Throughout the order interspecific variation occurs
in sclerotization, pigmentation, and vestiture of fe-
male genitalia and is of potential value in systematic
studies, at least at the species level.

Internal genitalia of males

Probably the distal portion of the ejaculatory duct
is projected into the vulva during copulation but it
rarely, if ever, has sclerotic rigidity—an aedeagus.
Exceptions to this are especially apparent in the ge-
nus Enveja Navas (Fig. 50), and to a lesser degree in
most genera of Anisembiidae, and in some species
of Oligotomidae; but this is merely limited to sclero-
tization of the duct walls. In my figures, labelled

40 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

7th tergum 8th tergum
x {

7th 8th 1
sternum sternum 2nd

Ul
laterotergites Ist valvula

9th tergum

sternum
\

_ 10th tergum

— cercus - basipodite

_-- cercus

\
accessory gland aperture

FIGURE 37. Abdominal terminalia of adult female Oligotoma nigra Hagen (Oligotomidae). Upper, lateral aspect.

Lower, ventral aspect.

gonapophyses, they appear as a pair of rod-like struc-
tures fused beneath the duct’s apex, the orifice of
which is microspiculate in many species. The scle-
rotic portions of such structures are most apparent in
cleared, slide preparations.

The need for a well-developed intermittent or-
gan is lessened by the fact that copulatory union is
accomplished and prolonged by use of processes,
lobes and hooks on the ninth, tenth and eleventh
somites and, in many species by a clasping action of
the left cercus and/or its basipodite. These, the pri-
mary characters used in classification, must, how-
ever, be used with caution because of frequency of
convergence. For example, fusion of segments of
the left cercus occurs in many unrelated taxa and, in
some cases, as a variable within a species.

External genitalia of males
(“Terminalia’’)

Complex, often bewildering, male abdominal
terminalia distinctions, often complicated by
convergences, are fundamental characters in system-
atic studies.

Early in its penultimate instar a male’s abdomi-
nal apex is identical to that of other nymphal stages.
Accordingly, the tenth tergum is triangular and un-
modified; the eleventh (epiproct) is represented by a
small, rudimentary sclerite (EP) just beneath the acute
apex of the tenth; the ninth sternite is transversely
quadrate; the anus is flanked by large, triangular, con-
vex paraprocts (LPPT and RPPT); the cerci and their
fleshy basipodites (LCB and RCB) are symmetrical.
Later in the penultimate instar future changes are pre-
saged by distortions due to developments within; the
tenth tergum may enlarge toward its left side, the ninth
sternite (H) may develop a small medial lobe (HP),

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 4]

laterotergites 7th tergum
N

ve ‘
co 7th 8th
pleurites sternum sternum

pleurite ~~ _ Baa
pleurite

i
laterotergites

8th tergum

Ist valyula —valyula sternum

gonopore \

9th tergum 10th tergum

epiproct
cércus -
basipodite

2nd Oth

_- anus

epiproct

—--- cercus

accessory gland aperture

FIGURE 38. Abdominal terminalia of adult female Metoligotoma ingens Davis (Australembiidae). Upper, lateral aspect.

Lower, ventral aspect.

and its left cercus-basipodite often exhibits signs of
profound changes. The appearance of highly modi-
fied adult structures following ecdysis and reduction
of some nymphal structures, e.g., the right paraproct,
iS most interesting.

The terminalia usually are intricately developed
to insure prolonged copulation and perhaps owe such
complexity to sperm competition within the species
to improve sexual union and thereby assurance that
the contents of a particular male’s spermatophore will
have time to enter the spermatheca.

In many genera the basal margins of terga of sub-
terminal abdominal somites are extended forward as
especially large apodemes anchoring intertergal
muscles which elevate the terminalia during copula-
tion. Such males often move about outside of galler-
ies with the terminalia curled forward in the manner
of male scorpionflies and earwigs.

In Clothoda, the order’s most plesiomorphic ge-
nus (Frontispiece and Figs. 39-41), the terminalia
are almost perfectly symmetrical and basically simi-
lar to those of nymphs and females. In C. nobilis
(Gerst.), the most plesiomorphic species of the or-
der, the tenth tergum (10) is short, narrowly trans-
verse, and medially uncleft. The tergite’s caudal apex
is turned upward as a thin somewhat translucent
medial flap (MF). The medio-basal portion of the
tergite is non-setose, weakly sclerotized, shallowly
depressed and projected cephalad as an area I have
termed medial sclerite (MS). This is a neutral area
separating the more vaulted, sclerotic, setose, incipi-
ent hemitergites (10 L and 10 R) to which the cer-
cus-basipodite muscles are attached.

In Clothoda longicauda Ross (Figs. 39A, 40B),
a slightly more apomorphic species, the medial area
of the tergite is partially membranous, forming a
branched medial cleft which, in adult males of most

42 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

other species of the order, divides the tenth tergum
into well-defined hemitergites (10 L and 10 R), each
of which has a distinct copulatory function.

Viewed from the caudal aspect (Fig. 39A) it is
apparent that the medial flap (MF) of C. longicauda
bears a small sclerite on its ventral surface (perhaps
a vestige of the eleventh tergite). Also visible is a
fleshy lobe above the anus which apparently is the
epiproct (EP), a rudiment of somite eleven. This
conclusion is confirmed when one views the caudal
aspect of Archembia batesi (McLachlan) (Fig. 39B)
and notes that the ventral sclerite of the medial flap
(MF) has elongated and is extended onto the epiproct
which has become an extensive supra-anal pad (EP).
A transverse fold, or hinge, occurs where the epiproct
levels off.

In C. longicauda, indistinct, small lobes, visible
on either side of the medial flap (MF) appear to be
precursors of hemitergal processes (10 LP and 10 RP).
In Archembia batesi (Fig. 39B), these lobes are dis-
tinct processes (10 LP and 10 RP) while the medial
flap (MF) remains prominent. This condition pre-
vails in many genera, as exemplified by Dihybocercus
lunaris (Navas) (Fig. 39C). In most of these genera
the medial flap (MF) has rotated clockwise so as to
almost parallel the longitudinal axis of the medial
cleft. Incidentally, in Dihybocercus and other
Embiidae, there is a small pouch at the forward end
of the medial flap. It is likely that this end of MF
produces glandular secretions of significance during
copulation. This deserves investigation. However,
in at least one major section of Embudae, the flap
(MF) usually is reduced to a longitudinal, sclerotic
ridge, or it may have completely atrophied.

In all genera of Clothodidae, except Clothoda,
the medial flap (MF), or at least its caudal angle,
seems to assume the function of the right process
(10 RP), for there is no flap-like structure in the nor-
mal position of the medial flap. In these genera the
epiproct (EP) is a broad, supra-anal pad, often with
a narrow, but prominent, longitudinal sclerite. Such
conditions, especially that of at least portions of the
medial flap (MF) serving as a process, are charac-
teristic of Enveja, Oligotomidae, Teratembiidae and
other taxa. Interestingly, Chromatoclothoda nana
Ross is well on its way toward becoming oligotomoid
in terminalia structure (Ross, 1987:34).

If this tentative interpretation is correct, then at
least portions of the right tergal process (10 RP) are
analogous, not homologous. It is therefore possible

that a major taxonomic division of the order begins
within the family Clothodidae. It will be noted that
the longitudinal membranous area between the me-
dial flap (MF) and the right hemitergite (10 R) of
most Embiidae has become transverse in
Oligotomidae and Teratembiidae and partially to
completely separates MF + 10 RP from an often
much-reduced 10 R. This enables MF and 10 RP to
hinge directly ventrad, or even forward, beneath the
hypandrium (H) during copulation.

Because of observation limitations within silk
galleries, copulation is difficult to observe. However,
in a male specimen of Aposthonia (Oligotomidae) pre-
served in alcohol, copulatory positions of various
structures were fixed. In this specimen the prob-
ably-composite 10 RP (MF + 10 RP) is folded down
and forward completely beneath the hypandium (H).
The epiproct (EP) is also pulled down so that neither
of the two structures is visible from above. Appar-
ently contraction of strong inter-tergal muscles at-
tached to EP is the force that moves the composite
right tergal process (which apparently lacks muscles).
The specimen also exhibits 10 LP pressed against the
inner side of the basal segment of the left cercus with
its complex apex vertical and faced to the right. The
hypandrium process (HP), forming a rigid trough for
the gonopophysis, is directed upward, like an erect
human penis and is pressed against the sclerotic mar-
gins of 10 L and 10 LP.

Throughout the order, the left hemitergite (10 L)
is Well defined, its margins usually sclerotic and in-
flexed and its surface vaulted to provide especially
strong anchorage for large muscles serving the im-
portant clasper function of the left cercus, or its
basipodite.

The left hemitergite’s process (10 LP) often con-
sists of an inner talon and an outer flange which of-
ten is thin, or fleshy. In many species, however, the
outer flange is greatly reduced, or absent. The left
hemitergite’s process assumes many forms consis-
tent within a species and thus is especially useful in
systematic studies. It probably is the most important
structure for providing rigid guidance of the apex of
the endophallus into the vulva.

Often the left hemitergite (10 L) is clearly sepa-
rated from other portions of the tenth tergite by a
submedial, membranous cleft which may extend to
the basal margin of the tergite. In many other spe-
cies, however, the basal portions of the cleft is ab-
sent due to fusion of the inner-basal margin of 10 L

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 43

medial flap MF
“| epiproct EP

-10 RP
~ right hemitergite

1OL JOR
left right
cercus - basipodite \: cercus - basipodite
LCB

RCB

‘cercus socket
left paraproct - ~~

~right paraproct
LPPT ght parap'

RPPT

left hemitergite _

P 10R
epiproct EP
ge
S
LCB -- --RCB
~ cercus
socket

left. paraproct ~~
LPPT

= right paraproct
RPPT

left hemitergite — _ tight hemitergite

right cercus

4
right paraproct

/
left paraproct hypandrium process

Cc

FIGURE 39. A. Caudal aspect of male terminalia of Clothoda longicauda Ross (Clothodidae). B. Caudal aspect of
Archembia batesi McL.) (Embiidae). C. Caudal aspect of Dihybocercus lunaris (Navas) (Embiidae). These drawings
show increasing complexity of the terminalia.

44 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

1
HP

LPPT RPPT

FIGURE 40. A. Dorsal aspect of male terminalia of Clothoda
nobilis (Gerst.), the most plesiomorphic species of the or-
der. B. Dorsal aspect of terminalia of Clothoda longicauda
Ross, a slightly more aposematic species of Clothodidae.

with the medial sclerite (MS), as in most
Teratembiidae (Figs. 52, 53). The medial sclerite
(MS) often is obscure, or absent, but in Teratembiidae
it is extensive, fused to the inner-base of 10 L and
usually projects (often acute in form) beneath the
entire left half of the ninth tergite (9). The fold of its
left side is continuous with that of the outer side of
the left hemitergite (10 L) and thus provides espe-
cially strong muscle anchorage.

The ninth sternite, or hypandrium (H), is a broad,
quadrate, subgenital plate which usually has a weak
basal margin but often has strong lateral margins. In
plesiomorphic species it is symmetrically produced
medially as a caudal process (HP) which subtends
the apex of the endophallus. In apomorphic species
it often is angled leftward and complexly modified,
as in Dactylocerca Ross (Anisembiidae) (Fig. 47).
The basal, non-setose, sclerotic portion of the para-
procts (LPPT and RPPT) often are closely associat-
ed, or fused, with the caudal angles of H. Indeed, the
left paraproct (LPPT) may fuse to become the scle-
rotic, left-caudal margin of H and, in some species
of Teratembiidae, the hypandrium process (HP) is
completely atrophied and the left paraproct becomes
the sole subgenital support (Fig. 42).

Primitively, the paraprocts (LPPT and RPPT) are
equal in size and form. Each consists of a fleshy,
setose, distal portion flanking the anus, and a basal,
sclerotized, non-setose portion. In Archembia (Fig.
39B), the distal (caudal) portion of the paraproct may
be membranous and setose and may atrophy while
the basal portion may fuse with adjacent structures.
This figure also illustrates the beginning of leftward
asymmetry of the paraprocts.

The basal segment of the left cercus (LC,) may
be unlobed, as in clothodids, but more often it has a
prominent inner lobe bearing numerous, small, conate,
peg-like setae (“echinulations”) which enhance the
segment’s copulatory grip. As further improvement
of this “tool,” especially in some Anisembiidae, the
distal segment is “absorbed” into the basal to form
an unjointed clasper. The extreme example is in
Dactylocerca Ross in which the segment is long, ar-
cuate and “embraces” the females left side (Fig. 47).
Such composite left cerci have independently devel-
oped on many unrelated evolutionary lines.

On a distinct evolutionary tangent, especially in

Teratembiidae, the copulatory grip is performed by
the extreme base of the basal segment, perhaps by

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 45

10th _- medial flap MF

tergite

ath —..11th tergite EP

tergite

9th
tergite right paraproct
RPPT
_tight

left LPPT -\_ cercus
, _ Paraproct
8th

sternite

2
-
-
-

7 left cercus LC,

hypandrium H

. left
9th sternite cercus-basipodite LCB

FiGure 41. Lateral aspect of terminalia of Clothoda nobilis showing upturned apex of the tenth tergite which becomes the
medial flap (MF), and the rudimentary eleventh tergite which becomes part of the epiproct (EP).

Mf,

iter
/

MM,

=
\9
uv
4

Vip

FiGuRE 42. Atrophy of hypandrium process (HP) as the ventral support of the ejaculatory duct and assumption of this
function by the left paraproct (LPPT) in males of two new genera of Teratembiidae. A. New species from Kenya with HP
still serving as ventral support. B. New species from India. C. New species from Nigeria. Also shown in Figs. 51-52.

46 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

the left cercus basipodite (LCB) which has varied
mesal processes (Figs. 52, 53).

The right cercus of males rarely is modified as it
apparently has limited, or no function, in clasping
the female and, in many cases, at least the basal seg-
ment is partially or entirely desclerotized. In very
few species, however, the inner face of the basal seg-
ment is sclerotic and even more rarely distally in-
wardly lobed. Such lobes never are echinulate. In
the Australian family Australembiidae the basal seg-
ment always is globular.

There are probably many other factors and
structures prolonging copulatory union. For example,
males often grasp the female’s head with highly
modified mandibles. Dense, large setae on the

FIGURE 43A. Anomalous terminalia of Oligotoma
greeniana Enderlein of India, “mirror-image.”

1OR+10R

hemitergites and inner sides of the cerci, as in
Pachylembia Ross, may assist. The reader should
refer to Figures 44-53, a “portfolio” of terminalia
figures at the close of this section, which show some
of the diversity of terminalia within the order.

Anomalous male terminalia

A small percentage of males have anomalous,
*miurror-image,” terminalia in which normally devel-
oped structures are completely reversed left-to-
right(Fig. 43A). In other anomalous specimens, struc-
tures of the left side are symmetrically repeated on
the right side (Fig. 43B), or those of the right are
repeated on the left (Fig. 43C). There also are occa-
sional bilateral gynandromorphs. These are conspic-
uous in the case of winged species—the female side,
of course, being wingless.

FIGURE 43B. Anomalous Aposthonia minuscula
(Enderlein) of India. Left side repeated on right side.

FIGURE 43C. Anomalous Diradius plaumanni (Ross) of
S. Brazil. Right side repeated on left side.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 1 49

ana
Ny Re
%
2 uP

FiGuRE 48. Metoligotoma illawarrae Davis Austral- FIGURE 49. Embonycha interrupta Navas (Embonychidae).
embiidae). Eastern Australia. Chapa, northern Vietnam.

50 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FIGURE 50. Enveja bequaerti Navas. Central Africa. FIGURE 51. Oligotoma nigra Hagen (Oligotomidae).
Middle East, introduced into southwestern USA and Aus-
tralia.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART | 5]

“GY 10 LP

J

LCB {/ 10 RP

FIGURE 52. Oligembia capensis Ross (Teratembiidae). FIGURE 53. Paroligembia n. sp. (Teratembiidae). Ethio-
Cape Region, Baja California, Mexico. pian highlands.

52 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

Literature Cited

Alberti, V. G. and V. Storch. 1976. Transmissions und
rasterelaktronenmikorshkopische Untersuchung
der Spinndrusen von Embien (Embioptera, In-
secta). Zool. Anz. (Jena) 197:179-186.

Barlet, J. 1985. Le Pterothorax du Male d’Embia
surcoufi Navas (Insectes, Embiopteres). Bull.
Soc. Royal Sci. Liege 54:349-360.

Barth, R., 1954. Untersuchungen an den Tarsaldrusen
von Embolyntha batesi McLachlan, 1877
(Embioidea) Zool. Jahrb. (Anatomie) Jena
74:172-188, 22 figs.

Bitsch, J. and S. Ramond, 1970. Etude du squelette
et de la musculature prothoraciques d’Embia
ramburi R.-K. (Insecta: Embioptera). Com-
paraison avec le structure du prothorax des autres
Polynéopteres et des Aptérygotes. Zool. Jahrb.
(Anatomie) 87:63—93, 12 figs.

Carpenter, F. M. 1950. The lower Permian insects of
Kansas. Part 10. The order Protorthoptera: the
family Liomopteridae and its relations. Proc.
Amer. Acad. Arts Sci. 78:185—219, 11 figs., 2 pl.

Carpenter, F. M. 1976. The Lower Permian Insecta
of Kansas. Part 12 Protorthoptera (continued),
Neuroptera, additional Paleodictyoptera, and
families of uncertain position. Psyche 83:336—
376, 23 figs.

Cockerell, T. D. A. 1908. Descriptions of Tertiary

Comstock, J.H. 1918. The Wings of Insects.
Comstock Publishing Co., Ithaca, New York.

Crampton, G.C. 1926. A comparison of the neck
and prothoracic sclerites through the orders of
insects from the standpoint of phylogeny. Trans.
Amer. Entomol. Soc. 52:199—248, pls. 10-17.

Davis, C. 1936. Studies in Australian Embioptera,
part | Systematics, part I]. Further notes on sys-
tematics. Proc. Linnean Soc. New South Wales
61:229-258, 50 figs., pl. XI.

Davis, C. 1939. Taxonomic notes on the order
Embioptera III. The genus Burmitembia
Cockerell. Proc. Linnean Soc. New South Wales
64:369-373, fig.

Enderlein, G. 1909. Die Klassifikation der Embiidinen,
nebst morphologischen und physiologischen
Bemerkungen besonders tiber das Spinnen
derselben. Zool. Anz. Leipzig 35:166-191.

Enderlein, G. 1912. Embiidinen. Coll. Zool. Selys
Long-champs, Cat. Syst. et Descr. Fasc. II, 76
figs, 4 plates.

Hagen, H. A. 1861. Synopsis of the Neuroptera of
North America, with a list of the South Ameri-
can species. Smithsonian Misc. Coll. 4(1):xx +
347 pp. (Citation as 1862 is erroneous).

Henning, W. 1981. Insect Phylogeny, American Edi-
tion, John Wiley and Sons, Chichester, N.Y.,
Brisbane, Toronto. xix + 514 pp., 143 figs.

Hong and Wang. 1993. Acta Palaeontologica Sinica
32:141-150.

Hurd, P. D., R. F Smith, J. W. Durham. 1962. The
Fossilferous amber of Chiapas, México. Ciencia
(Méx.) 21:107-118.

Imms, A. D. 1913. Contributions to a knowledge of
the structure and biology of some Indian insects,
I. On Embia major sp. nov. from the Himalayas.
Trans. Linnean Soc. London (Zool.) (2)11:167—
195, 6 figs. 3 pl.

Kristensen, N. P. 1975. The phylogeny of hexapod
“orders.” A critical review of recent accounts.
Zeitschrift Zool. Systematik Evolutions-
forschung 13(1):1-44, 8 figs.

Matsuda, R. 1960. Morphology of the pleuro-ster-
nal region of the pterothorax in insects. Ann.
Entomol. Soc. Amer. 53:712—731, 38 figs.

Matsuda, R. 1965. Morphology and evolution of the
insect head. Mem. Amer. Entomol. Institute
53:1-334.

Matsuda, R. 1970. Morphology and evolution of the
insect thorax. Mem. Entomol. Soc. Canada
76:1-431, 172 figs. (pp. 132-138, figs. 49-51).

Matsuda, R. 1976. Morphology and evolution of the
insect abdomen. Pergamon Press, Oxford and
New York. viii + 532 pp., 155 figs. (pp. 158-160,
fig. 34.

Melander, A. L. 1902. Two new Embiidae. Biol. Bull.
3:16—26, 4 figs.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART | 53

Mukerji. 1928. On the morphology and bionomics
of Embia minor n. sp., with special reference to
its spinning organ. Rec. Indian Mus., Calcutta
29:253—282, 10 figs., 1 pl.

Pictet, F. J. 1854. Traité de Paleontogie, ou Histoire
naturelle des animaux fossiles. Paris, J. B.
Bailliere, 2nd ed., Vol II, 10 pls.

Rahle, W. 1970. Untersuchungen an Kopf und Pro-
thorax von Embia ramburi Rimsky-Korsakov
1906 (Embioptera, Embiidae). Zool. Jahrb.
Anatomie 87:248—330.

Rimsky-Korsakov, M. 1914. Uber den Bau und die
Entwicklung des Spinnapparatus bei Embien.
Zeit. Wiss. Zool., Leipzig 108:499-519.

Ross, E. S. 1956. Anew genus of Embioptera from
Baltic Amber. Mitt. Geol. Staatinst. Hamburg
25:76-81, 2 figs.

Ross, E.S. 1987. Studies in the insect order
Embiidina: A revision of the family Clothodidae.
Proc. California Acad. Sci. 45:9—34, 12 figs.

Slifer, E. H. and S. S. Sekhon. 1973. Sense organs
on the antennal Flagellum of two Species of
Embioptera (Insecta). Jour. Morph. 139:211—
216, 1 fig., 3 pls. (pp. 218-225).

Snodgrass, R. E. 1935. Principles of Insect Mor-
phology. McGraw-Hill Book Co., New York and
London. ix + 667 pp.

Storozhenko, S. Y. 1997. Fossil history and phylog-
eny of orthopteroid insects. Chapter 4 in The
Bionomics of Grasshoppers, Katydids and their
Kin. S. K. Gangwere, M. C. Muralirangan and
M. Muralirangan editors. CAB International,
Oxon, England and New York. xiii + 529 pp.
(Chapter 4, pp. 59-82).

Szumik, C. A. 1994. Oligembia vetusta, a new fossil
teratembiid (Embioptera) from Dominican Am-
ber. Jour. New York Entomol. Soc. (102)1:69-73.

Zeuner, F. E. 1936. Das erste Protoperlar aus
Europaischem Perm und die Abstammung der
Embien. Jahrb. Geol. Landesanstalt 56:266-273.

al euia
2 Obit Wiliae ;

Part 2
A Review of the Biology of Embiidina

Summary

The biology of Embiidina is reviewed and
illustrated with many of my photographs. Stressed
are the evolutionary restrictions imposed by life
almost completely confined to self-produced, narrow,
silk galleries. This review also covers diverse topics,
such as: diet, locomotion, social behavior, mating,
eggs and their protection, development, ecological
and geographic ranges, natural enemies, and
diseases.

In writing EMBIA Part 1, on anatomy of the
order, it was necessary to discuss the relationship of
structure to function, especially in reference to the
wings. Therefore, the reader must expect some
repetition of information in the two parts.

Methods

To secure specimens for a comprehensive, world
scope coverage of embiid taxa, | made many, often
extensive, collecting trips throughout the order’s
range. For example, almost all countries of Africa
were visited during eight excursions covering about
four years in all. Nine months were spent in India,
Bangladesh and Pakistan, several months in
southeastern Asia, Australia and many months in
significant regions of the Americas, and other places.

To be effective, and to avoid wasting time and
tunds in hotels, major trips were made in personally-
designed camping vehicles (see National Geographic
articles in March 1961 and September 1965 issues).
To get a broad representation of higher taxa,
vegetation and life zone maps were used to determine
routes of travel.

Such extensive fieldwork over a fifty-year period,
offered opportunities to observe embiid biology, but
only briefly, for life histories are often a year in
length. Prolonged observations had to be made in
cultures maintained in my Academy and home
laboratories (Fig. 43). However, because as many
as 800 cultures resulted from a single eighteen-month
expedition, it wasn’t possible to make an in-depth
study of any one species.

General biology

In spite of presumed great antiquity and isolation
of taxa on long-separated continents, the biology of
embiids is remarkably uniform, as it is in several
other ancient arthropod groups, such as scorpions
and cockroaches. In embiids, order-defining
characters and biological uniformity result from
perfection of evasive movement in a self-produced
micro-environment, one which literally and
figuratively has “channeled,” or limited
diversification of the order’s anatomy and biology.
The key factor, of course, is life almost entirely
confined to narrow silk galleries (Figs. 1, 2 and 3).

The galleries are produced by unique foretarsi
swollen by numerous, perhaps hundreds, of globular,
blastula-like, syncytial glands within the basal
segment (Fig. 2). Viscous silk is conducted from
each gland via a narrow duct to an opening at the
tip of a hollow, seta-like silk-ejector. These are
located mostly on the thin, ventral surface of the basal
segment and to a much lesser extent on the mid-
tarsal segment. In Oligotoma nigra Hagen, for
example, there are approximately 150 such ejectors
on each tarsus and thus a corresponding number of
silk strands may simultaneously issue with each
tarsal stroke. As both legs spin, silk webbing is
produced with remarkable rapidity. Indeed,
considering production speed and quantity, embiids
may rival spiders as the most efficient silk-producing
organisms on Earth.

Except perhaps for their number, it is assumed
that the glands are similar in all developmental stages
of all species of the order. Even first instar and
teneral individuals are able to spin and, remarkably,
the ability continues throughout adult life.

The galleries compose an expanding labyrinth
usually produced and occupied by the brood of a par-
ent female. In some species, however, it is neces-
sary for early stage nymphs to disperse and estab-
lish independent galleries so as to avoid injury, or
death, due to sibling hostility. It should be noted,
however, that such hazards are likely to be intensi-
fied in crowded laboratory cultures in which most of
my observations were made.

OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FIGURE 1. Adult Pararhagadochir birabeni (Navas) (Embiidae) of Argentina, showing typical ap-
pearance and posture of all female embiids. Pale bands between thoracic somites characterize many
species of the order.

FIGURE 2. Adult female of Ptilocerembia n. sp. (Notoligotomidae) from Malaya exhibits universal
spinning foretarsi, large hind femora and short, two segmented cerci. The vulva opens between the
dark, sclerotic eighth and ninth abdominal sterna.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 3)

FIGURE 3. Nymphs radiate in galleries, often used in common, which fit their body size. A
portion of the parent’s larger gallery crosses the top of this photograph, Donaconethis n. sp.

(Embiidae), Eritrea.

Embiids are highly thigmotactic and produce gal-
leries just narrow enough to maintain constant con-
tact of body vestiture with walls. Advantageous
confinement within galleries apparently has governed
anatomical and behavioral trends over a long period
of evolutionary time. Thus, the spinning organs ap-
pear to antedate and to have regulated the evolution
of all order-defining characters. Such characters must
have become fixed in Upper Paleozoic or Lower Me-
sozoic times and fully distributed within the order
before the breakup of Pangaea.

Gallery diameter correlates with the size of the
embuid frequenting the particular section of a colony.
In most species first and second instar nymphs, after
clustering for several days near their mother (Fig. 4)
commence to radiate outward in their own galleries
which are increased in diameter and extent as the
maker grows. Because offspring of most species hatch
from a single egg mass and develop in unison, the
galleries of a brood usually become interconnected
and used in common.

In addition to the tarsal silk glands, ordinal char-
acters include an elongate, supple body; a prognathus

FIGURE 4. First and second instar nymphs usually clus-
ter near the parent female, Embia sp.

OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FIGURES 5 and 6. Especially in arid regions, parent females often find refuge in rock crevices. Follow-

ing rains, galleries of the brood radiate on the rock’s surface as lichens are “grazed.” During excessive
heat, or bush fires, the embiids crowd back into the crevice. Scelembia n. sp., (Embiidae), Angola.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 5

head with a sclerotic, ventral bridge; short legs, effi-
cient backward movement; equal wings with reduced
vein strength except for sinus veins which temporari-
ly can be stiffened with blood pressure; tactile, two-
segmented cerci; universal neoteny in females,
partially or entirely so in males of some species.

These, and other features treated elsewhere in this
work, maximize survival almost entirely dependent
on confinement in narrow galleries. Such galleries
are constantly extended within or on edible surfaces.
Therefore, embiids can feed without leaving their
galleries. Food is very simple. All species are pri-
marily phyto-scavengers but some may also occasion-
ally eat live mosses, lichens, etc. There are no
predatory species.

Some researchers have suggested that the galler-
ies control humidity, others, including me, have con-
cluded that their primary function 1s protection from
predators and parasites. Although it seems unlikely
that the galleries can regulate humidity to any appre-
ciable extent, I have observed the dense gallery walls
are at least temporarily impervious to water and thus
may protect the occupants, especially soil-inhabiting
species, from short-term flooding or habitat satura-
tion after heavy rainfall.

Embiids, such as Haploembia spp.. inhabiting
regions with cold periods, even those temporarily
blanketed in snow, completely enclose themselves in
a cocoon spun within the galleries. [have noted them
in the introduced species, Haploembia solieri (Ram-
bur), in California, as well as in Mediterranean and
Turkish regions. The habit may be widespread in
species living in seasonally-cold regions. As with
cocoons of moths, and other insects, the enclosures
must function primarily as predator barriers. This is
especially important to embiids in seasonally-cold
regions because potential enemies, such as predaceous
beetles, may be able to hunt prey at low temperatures
which immobilize embiids.

The primary advantage of gallery life seems to be
predator-avoidance and this is increased when gal-
leries extend beneath or within, solid objects (Figs.
5, 6). Protection in exposed galleries may be some-
what indirect because the silk isn’t strong enough to
wall off most predators. A predator’s initial contact
with a web surface probably broadcasts a tactile warn-
ing which stimulates rapid, usually reverse, move-
ment into a deeper, more rigid recess within the
labyrinth. Edgerly’s study (1997) of ant entrance

holes in gallery walls, based on Antipaluria urichi in
Trinidad, appears to contradict this conclusion, at
least in her study area. While collecting embiids
throughout the order’s geographic and ecological
range, I have frequently encountered embiid colo-
nies in direct contact with those of ants, often under
the same stone, without any apparent molestation by
the ants. Ants, however, are probably the principal
predators of embiids whenever they leave their gal-
leries.

The value of rapid silk web production was ap-
preciated by me when, experimentally, I released em-
bids on tropical tree trunks. An exposed embiid
immediately retreats into the nearest bark crevice and
at once begins to cover it with a silk web. This barri-
er is steadily improved and extended and, if favor-
ably located, may become the locus of a new labyrinth.
It is more likely, however, that the exposed embiid
will immediately be seized by an ant or other preda-
tor before it can get into a crevice and spin a barrier.
It may be said that an embiid outside of its gallery is
almost as much out of its element as a fish out of
water.

Life in silk galleries may offer other benefits. It
is probable that embiids conserve energy by having
pre-constructed, smooth-surfaced runways to and
from a food source. Silk itself may be an excretory
by-product put to good physiological use. There is
also the advantage of being able to silk-partition fe-
cal pellets from within the galleries, thereby main-
taining debris-free avenues of movement. After
excreting a pellet, an embiid snips a hole in the adja-
cent gallery wall and, using its mandibles as tongs,
places the pellet outside and then closes the opening
with silk. Fecal pellets accumulated between galler-
ies tend to strengthen their walls (Fig. 7A) and, when
deliberately placed atop a labyrinth’s surface on a
tropical tree trunk, serve as a medium for growth ofa
microflora enhancing the cover. The habit of pro-
tecting, or not protecting, exposed galleries with a
surface covering may characterize a species, or even
a genus. In addition to fecal pellets, some species
pulverize outer bark and cause gallery surfaces to
become dusted with chewed powdery debris which
may almost completely conceal the colony (Figs. 8,
9). In contrast, species of many other genera never
cover their galleries and thus they can be seen from a
considerable distance (Figs. 25, 26).

Another type of feces disposal involves their ac-
cumulation in low mounds here and there on the la-

6 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FIGURE 7A. Embiids place their dry fecal pellets outside
of their galleries, thereby insuring debris-free avenues of
movement and a strengthening of gallery walls. Archembia
batesi (McL.) (Embiidae) in an Amazon rainforest. Sur-
face layer of silk removed.

Qe
et

OQ, *

FIGURE 7B

ously white and, in this case, aligned with bark crevices

Galleries of Archembia batesi are conspicu-

FIGURE 8. In contrast to Archembia, many embiids, as a
generic habit, deposit feces on exposed surfaces of their
galleries. Chromatoclothoda n. sp., ( Clothodidae). Ec-
uadorian montana.

FIGURE 9. Galleries of anew genus and species from south-
eastern Asia’s Golden Triangle, are completely concealed
beneath pulverized bark particles and feces. Doi Pue, Thai-
land.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 7

byrinth’s substrate. These are then progressively
covered with layers of silk and appear as low white
mounds similar to those covering egg masses of some
species (Fig. 42).

In tropical forest habitats another type of protec-
tion involves prolonged use of old galleries as a sur-
face cover over those spun beneath by successive
broods. Thus they can function somewhat as a layer
of bark. Galleries themselves may also have pro-
longed use and consequently silk becomes increas-

walls through flexibility (vein atrophy) and alternat-
ing stiffening for flight by blood pressure in sinus
veins. The ultimate accommodation is wing size re-
duction, or complete loss, through neoteny (or pae-
domorphosis).

These wing specializations are so complex and
universal that it is inconceivable that they evolved
solely to increase survivability of adult males. In-
stead, they must have developed ages ago when the
order was confined to tropical zones of Pangaea as a

FIGURE 10. Adult male during defensive backward movement. Temporarily flexible wings bend forward, thereby reduc-
ing friction which could slow escape. Actually, wings when not used in flight can bend at any point—even crumple. New

genus and species of Oligotomidae, Thailand.

ingly dense and obscured by surface debris. Thus, in
Amazonia, galleries of species of Chromatoclotho-
da Ross (Fig. 8), in contrast to the conspicuous, white
galleries of more plesiomorphic Clothoda Enderlein
(Fig. 23), both family Clothodidae, can be located
only by random tweezer-scraping of likely surfaces,
such as the underside of laterally projected branches
or ledge overhangs.

The presence or absence of wings and their pecu-
liarities are directly related to predator-avoiding re-
verse movement in silk galleries—the need to reduce
or overcome wing friction, or snag, against gallery

means of increasing predator-avoidance by adult fe-
males which must live long enough to produce and
guard eggs and early instar broods. In contrast, males
are short-lived and contribute only sperm to the re-
productive process. Later, however, more effective
reverse locomotion of females was achieved through
complete apterism by means of neoteny. In a sense,
females of all extant species are second or third in-
star reproductive nymphs increased in size. Although
males of most species possess wings, there is a ten-
dency on almost all evolutionary lines for them to
become completely apterous or brachypterous
through neoteny.

8 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

Embiid wings are non-deciduous, nearly equal in
size and shape, wide spaced in thoracic attachment,
flat and unfolded when in repose over the dorsum,
have pigmented stripes separated by hyaline inter-
vals following the courses of all longitudinal veins,
and, most important—indeed unique—some veins are
broad, glossy, cuticularized, longitudinal, blood si-
nuses. Except for slight anal (vannal) lobing in wings
of certain species of the rather plesiomorphic genus
Archembia Ross, the anal area of embiid wings is
greatly reduced. Such reduction even occurs in
Clothoda Enderlein, the order’s most plesiomorphic
genus. It is probable, however, that embiids had an-
cient ancestors with a well-developed anal area in the
hindwings comparable to that of Plecoptera and Mas-
totermes in Isoptera.

Thus the principal evolutionary trend in embiid
wings wasn’t improvement for flight but, instead, to-
ward rendering their possession less of a handicap
during movement within galleries. The ultimate ac-
commodation is complete aptery of all females as well
as of males of many taxa. A similar disadvantage of
wing possession was faced by sexual termites adapt-
ing toward easier movement in galleries in earth and
wood. In this case, however, the disadvantage 1s elim-
inated by wing break-off by nuptial adults prior to
copulation. Also, lifelong aptery of most individuals
in a termite colony is caused by endocrinal retarda-
tion of the appearance of adult structures and func-
tions (neoteny). A similar retardation is probably
responsible for the universal aptery of adult female
embtids, as well as varying degrees of brachyptery
to complete aptery in males of many embiid species,
genera, and even an entire family (Australembiidae).

The trend toward apterism in males is presently
active and has been so for at least the entire Tertiary
period, as evidenced by complete apterism of males
of Electroembia antiqua (Pictet), Baltic Amber
(Eocene?). Degrees of male aptery and brachyptery
occur on most evolutionary lines, such as: (1) males
with robust (nymphoid) bodies and short wings; (2)
males with wing pads similar to those of various on-
togenic stages from buds (gemmae) to full pads; (3)
complete aptery without even traces of wing buds.
In some species, such as Anisembia texana (Mel.), a
percentage of adult males in a population have nor-
mal wings, but most possess only tiny wing buds. In
Oklahoma, at the northern range of A. texana, all
males are completely apterous without traces of wing
buds.

Because females are universally apterous, flight
of males cannot increase geographic range of a spe-
cies, or enable a population to move away from envi-
ronments uninhabitable as a result of sudden or
gradual adverse ecological changes. Range exten-
sion and relocation can be effected only by females
surviving hazardous movement afoot outside of their
protective galleries, or by being carried in materials
transported by wind, water or human commerce. Male
flight, however, fosters random mating and thereby
reduces potentially disadvantageous incestuous mat-
ings so likely in gregarious, subsocial populations.
The complex subject of embiid wings is more fully
treated in my review of the order’s anatomy (EM-
BIA Part 1).

Social behavior

A typical embiid colony is a “gynopaedium”—a
parent female and her brood living together. Often
galleries of broods of adjacent females become in-
terconnected and the nymphs intermingle without
hostility.

Although there is no evidence of a division of la-
bor, or castes, some social advantages could result
from utilization, by some species, of preexisting gal-
leries produced by previous generations which had
occupied the same bark surface. In rainforests a mat
of such galleries may thus serve like a layer of bark
protecting new galleries spun beneath by succeeding
generations of nymphs. However, to reach ungrazed
edible surfaces, most species produce new labyrinths
radiating out from such initial coverings. Incidental-
ly, the silk of new galleries of some species is laven-
der in color.

The most important social activity is guarding
eggs and young by parent females in a manner simi-
lar to that of Dermaptera (Figs. 18, 42) (Edgerly, 1987a
and b; 1988). Early instar nymphs usually congre-
gate near their mother and perhaps benefit from her
presence for at least two instars (Fig. 4). As I service
laboratory cultures, young are often inadvertently dis-
associated from their parent but, in a short time man-
age to reassemble in spite of the disadvantage of hav-
ing to spin new galleries to do so. An aggregation
pheromone may be involved in this.

Food provisioning in arid regions by subterra-
nean species can also be regarded as social activity
for it tends to insure dependable, readily accessible,
food for the brood. It also avoids energy loss and

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 9

risks consequential to making repeated forays to an
often inclement and hazardous surface environment.

Embiids may be termed subsocial communal in-
sects. In this category females remain with their off-
spring for a period of time and members of the same
generation use the same composite nest without co-
operating in brood care. Such an interaction is an
inevitable consequence of gallery life. A lone female
simply lays eggs in a favorable place—often within
existing galleries, or in a new site, and her offspring
having no need to disperse, extend individual galler-
ies no farther than needed to reach food. Upon matu-
rity, her offpspring do not need to disperse to find
food or a mate. Therefore, females are likely to de-
posit eggs not far from their own place of origin.

In communal species, especially those occupying
tree trunks, or rock and road bank surfaces in the trop-
ics, a colony may grow to great size and even enve-
lope a huge tree trunk (Fig. 26). Theoretically, such
growth is radial—expanding at the periphery as edi-
ble surfaces are sought. In jar cultures and under
other artificial conditions, colony growth is three-di-
mensional, as reported by Friederichs (1913) in ref-
erence to an outbreak of Aposthonia gurneyi
(Froggatt) in a sugar refinery in Australia, or of Olig-
otoma saundersii (Westwood) in piles of stored pea-
nuts in Senegal.

There are exceptions to a gregarious habit, how-
ever. In some species, especially those found in the
savanna woodlands of central Africa (e.g., Dinembia
Davis spp.), nymphs are intolerant to one another and
must disperse soon after hatching and develop in in-
dividual galleries to avoid injury.

It is conceivable that if ever a form of reciprocal
or proctodeal feeding, or body licking should evolve
which would permit transmission of maturity-inhib-
iting pheromones, a worker caste might develop in
embiids. There is, however, absolutely no indication
or need of such behavior in Embiidina. Food ex-
change between embiids has never been observed and
proctodaeal feeding potentials are lessened because
excrement consists of dry pellets which are deliber-
ately placed or partitioned outside of the galleries im-
mediately following defecation. Nymphs hatching
from eggs don’t even eat the hardened pulverized ma-
terial, which in part may be fecal, placed around the
eggs by females of most species. Covering and side-
by-side placement of eggs appears to reduce ovipo-
sition by parasitoid wasps (Figs. 19, 42).

Female embiids could be likened to nymphoid re-
productives in termites, but it is most likely that their
neotenization is programmed by regulation of juve-
nile hormone rather than exchanges between individ-
uals.

Behavior of adult males

Upon maturity an adult male usually remains in-
active for a few days in the gallery section where fi-
nal ecdysis took place (Fig. 12). During this time its
derm hardens and becomes fully melanized, or pig-
mented. Concurrently, its nymphal pelt slowly pass-
es through the gut and is excreted. Later a male may
wander about within the galleries and may mate with
a receptive female, perhaps a sister, particularly in
species with apterous males which are more likely to
remain in a colony with sisters. Sister-mating is like-
ly in laboratory cultures and in gregarious species but,
in some species, males and females develop in sepa-
rate galleries. In such cases a male must vacate his
“personal” gallery, locate and bite his way into one
occupied by a female (Fig. 15).

In cultures, usually during warm afternoons, adults
of both sexes often move to uppermost levels of a
culture, protrude their forebodies from gallery open-
ings (Fig. 14) and, with their heads often hypogna-
thously angled, they rapidly vibrate their antennae.
It is assumed that such activity encourages sexual
contact. No investigations have been made to deter-
mine if “calling” pheromones are released during
such exposure, but this is likely.

Under certain meteorological conditions, often
just after the first rains ending a dry season, adult
males usually leave their galleries and take flight (Fig.
11), or, if apterous, simply run about on the ground.
In arid regions males of pale species have large,
coarsely-faceted eyes and tend to fly during warm,
humid nights when many other insects, notably nup-
tial termites, also fly. During such periods collectors
should always search for male embiids attracted to
lights, not only to collect specimens, but also as a
means of determining what nocturnal species of the
order have colonies in the vicinity. At least a portion
of a light sheet should be on the ground inasmuch as
males of some pale, nocturnal species are apterous
and can only run to lights. In some species, such as
Aposthonia tillyardi (Davis) of western Australia,
apterous and alate males may occur ina single local-

ity.

10 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

Color of adult males

Adult males of most species are melanized, at
times with a reddish prothorax, and/or a golden
pterothorax and metallic blue wing veins and sheen
(Fig. 13). Dispersal of colorful males is diurnal and
one may see alate males of such species in flight, or
resting on vegetation but, more likely, yet rarely, they
will be collected by random sweeping. Probably
mortality of diurnally dispersing males is very high
due to increased exposure to birds, or the elements.

Some protection against predation may involve Ba-
tesian mimetic resemblance to chemically-protected
diurnal beetles and stinging ants.

In various, usually unrelated taxa, the distal an-
tennal segments may be abruptly white, as also are
one or both cercus segments. In many species pale
intersomatic thoracic bands are present, as well as
longitudinal, pale, pleural abdominal stripes. All such
characteristics are due to white fat visible through a
transparent integument.

FiGurE 11. Alert posture of an adult male about to take flight. The strongly cuticular-
ized wing veins of this genus are prominent in this photograph. Enveja bequaerti Navas.

Katanga, central Africa.

FIGURE 12. This teneral male will remain for a long time in one place until fully hardened.
Incidentally, it has especially large wings, a characteristic of high altitude species.
Pararhagadochir n. sp., (Embiidae). Machu Picchu, Peru.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 1]

FicurE 13. Adult female and penultimate male of new
genus and species of Oligotomidae from Thailand. The
prothorax of both sexes of this dark brown species is bright
orange. Body length of female 18 mm.

Mating

When a male locates a gallery containing a re-
ceptive female, he bites an opening (Fig. 15), enters,
and approaches the female head-on, rapidly jerking
his body and vibrating his antennae. If the female is
unreceptive, her reaction, perhaps varying according
to species, may be antagonistic and dangerous. In
some encounters a female may attempt to eat, or at
least bite an approaching male. She lunges toward
the male with the same motions used in defending
eggs, or young brood. Often there is antagonism or
fighting between males.

When a female is receptive, there is mutual quiv-
ering of the antennae, head and prothorax and, alter-
nately, forward darting and retreat. Inasmuch as there
may remain a threat to the male, males of many spe-
cies reduce danger by grasping the fore portion of
the female’s head with the mandibles. Mandibles of
males exhibit varying degrees of specialization for
such a grasp, the most extreme of which are the large,
elongate, arcuate type characterizing some species of
Enveja Navas. Mandibles of males may also be used
to gently nibble the female’s body. Because the man-
dibles of adult males are not used for grinding food,

Ur

FIGURE 14. Epigamic females may protrude their forebodies
from galleries, possibly emitting pheromones to attract
males. “Embia” surcoufi Navas, (Embiidae). Mozambique.

FIGURE 15. Adult male (new genus and species, Oligotomi-
dae) biting an entrance into a gallery presumably occupied
by a receptive female. Thailand.

but primarily seem to be secondary sexual organs,
they have greatly diversified and are thus useful char-
acters in systematics. The short, robust, nymphoid
mandibles of adult females, used for food-grinding,
vary little throughout the order.

Usually, a male grips a female’s head across the
frontal region but there are variations. In one abnor-
mal case, I observed a male holding the dorsal cervi-
cal region instead of the head. Once secure in a
mandibular grip, the female’s head usually is pulled
to the right and the tip of the male’s abdomen probes
down her right side, thence leftward and upward be-
neath her genital opening. It is remarkable that a nght-
side approach is constant in all species of the order.
In any event, it explains the almost universal leftward
asymmetry of the male terminalia (to extend its
“reach”) and why the specialized left cercus functions
in many species as a clasper against the female’s left
side

Because mating occurs obscurely within galler-
ies, it is difficult to observe. However, Stefani (1953a,
c) made detailed descriptions of copulation in Embia
ramburi R. K., Cleomia guareschii Stefani, and Hap-
loembia solieri (Rambur). Earlier, Friederichs (1934)

ters, the male (left) lacks lobing on the left cercus and has
a greatly reduced left tergal process. The copulatory grip
appears to be increased by pressure of the dense, bristle-
like setae borne on the left and right hemitergites.

12 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

made brief observations of Embia ramburi and
Oligotoma nigra Hagen.

A few of my observations are recorded as fol-
lows:

(1) Archembia lacombea Ross (Embiidae). Bra-
zil: Rio de Janeiro. A male was observed rubbing
his submentum against the vertex of a female’s head.
In this species, and congeners, there are relatively
dense, often parallel, setal clumps, as well as foveae
on the submentum in and around which white secre-
tions collect. During this rubbing the male’s anten-
nae extended on either side of the female’s body. She
wriggled sinuously and continued a limited spinning
movement of her forelegs. Females of Archembia
Ross, and related genera, have a transverse, pale, of-
ten golden, eye-to-eye band above the brain. One
should investigate the function of this pale macula-
tion. Is it associated with mating, or does it have a
light-perception function?

(2) Machadoembia Ross, n. sp., (Embiidae)
Angola: near Quilenda.

Male grasped female’s head (face to face) with
his mandibles across her clypeus. The female fre-
quently lurched but the male maintained his grip for
at least a minute. During this time the male’s genita-
lia united with those of the female. These discon-
nected before the male released his grip on the head.
When freed the female walked off unharmed by the
male’s mandibular grip.

During other matings, males of this species
grasped heads of females from several frontal angles,
as well as the cervix. The female’s head was twisted
to the right as the male’s genitalia quickly sought con-
tact. In spite of much tugging, females seemed re-
ceptive to mating. One copulation lasted about 60
seconds.

(3) Parembia major (Imms), (Embiidae). India:
Mussourie U.P.

Male gripped female’s head frontally.

(4) Embia n. sp., (Embiidae). Ethiopia: Naza-
reth.

Male faced female and gripped her head with his
mandibles behind her eyes. The male twisted so that
most of his body paralleled the right side of the fe-
male and the tip of his abdomen crossed beneath the
female’s genitalia. Actual genital union wasn’t ob-
served.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 13

(5) Embia mauritanica Lucas (Embiidae). Alge-
ria: 27 mi. N. M’Sila.

Ventral concavity of median flap (MF) covered
inner curvature of left tergal process (10 LP). Pro-
cess of left paraproct (LPPT) pressed into dorsal de-
pression of left cercus lobe. Ventral nodule of LPPT
prevented left cercus lobe from moving ventrad.

(6) “Parembia” dobhali Ross, (Embiidae). India:
Dehra Dun.

Female began to eat copulating male while the
genitalia were sti!l joined.

(7) “Parembia” n. sp., (Embiidae). India: Bad-
amtan Forest Res. W. Bengal.

A male attempted to mate with another male
while holding its head in his jaws (behind the eyes)
and pushing his genitalia beneath the other male’s
wings, apparently mistaking this surface for that of
a female’s abdominal venter. Such abnormal ap-
proaches were observed on several occasions.

(8) Enveja bequaerti Navas, Zaire: 12 mi. S.

Sampwe.

Grip of head with jaws not observed but the re-
markably large mandibles suggest that they are
adapted for head-clasping. During one mating the
sexes remained parallel; the male on the right side
of the female with his terminalia angled leftward
and upward to join the female’s genitalia. They re-
mained united for about thirty minutes. Close ex-
amination revealed that the male’s right tergal process
was folded ventrad against the surface of its hypan-
drium and pressed against the female’s second valvi-
fer. After the pair separated, a hard, irregular,
gelatinous object, probably a spermatophore, pro-
truded from the vulva.

(9) Dactylocerca Ross, n. sp. (Anisembiidae),
Mexico: Alamos, Sonora.

The male’s long, arcuate, one-segmented, left cer-
cus embraced the left side of the female’s abdominal
apex from beneath. The grip was so tight that mem-
branes at the base of the female’s righ cercus be-
came distended. At no time did the heads connect
although this might have occurred before the obser-
vation began. Males of the genus have very small
mandibles.

(10) Australembia nodosa (Davis) (Australem-
biidae). Queensland: Millstream Falls. Mating
fixed in alcohol.

Male terminalia centered beneath female. The
apex of LC1+2 depressed membranes between left
basal corner of H and between caudal tips of pleu-
rite and laterotergite of somite 8. The hypandrium
process (HP) and tenth tergite (10 RP) pushed into
vulva. After separation, a spermatophore wasn’t
visible in the vulva. Perhaps the pair was killed in
alcohol before copulation was completed.

(11) New genus and species (Teratembiidae),
Transvaal: 18 mi. S. Louis Trichardt.

Mates faced the same direction. The male did
not grip female’s head and was somewhat beneath
her, his terminalia turned upward to the vulva, well
centered. Because it is unlobed, the left cercus did
not seem to be used as a clasper.

Later the female walked forward causing the two
insects to face opposite directions. The female con-
tinued to walk out of sight into a gallery dragging
the male backward as the genital union continued.

(12) Aposthonia Krauss, n. sp. (Oligotomidae).
Queensland: Brookdale (coastal! plain).

In a male specimen preserved in alcohol, the
right tergal process (10 RP) was folded down and
completely pressed against his hypandrium (H). The
epiproct (EP) was also pulled down and provided
musculature for movement of 10 RP. The left tergal
process (10 LP) paralleled the inner face of the bas-
al segment of the left cercus (LC1) with its complex
apex vertical, its dorsal surface facing toward right;
its left side was appressed on the inner apex of LC1.
The hypandrium process (HP) and the gonopophys-
is Were projected dorsad (almost vertical, like an erect
penis) and, pressed by the inner angle of the sclerot-
ic left hemitergite (10 L) and its process, provided
rigid enclosure for the ejaculatory duct.

These “mechanics” are probably universal in the
Oligotomidae, Teratembiidae, and other taxa which,
by convergence, have a transverse, membranous sep-
aration of EP and 10 RP from 10 R which serves as
a hinge permitting downward movement of 10 RP.

Eggs and their protection

Eggs of all species are remarkably similar (Fig.
17). They are tubular in form, basally rounded,
slightly curved, and have a large, slanted, strongly-
rimmed operculum. Their general appearance is sim-
ilar to that of bedbug eggs.

14 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

Eggs issue from the vulva with the operculum in-
ward and are deposited within the galleries, usually
attached to a substrate. However, in some species
they are loosely clustered in the galleries and not
imbedded in a hardened paste. In some species un-

FIGURE 17. All embiid eggs are similar in shape, have a
rimmed operculum, and are laid on their back. Haploem-
bia solieri (Rambur) (Oligotomidae) endemic to the Med-
iterranean region.

covered eggs may form a tunnel within which the
guarding female rests. Most often, however, eggs are
laid in a single-layered cluster and are imbedded in a
hardened paste of habitat material pulverized by the
female and deliberately placed as the eggs are laid
(Figs. 19, 20). It is probable that fecal pellets are
also pulverized for this purpose. I have counted more
than two hundred eggs in a cluster laid over a period
of several days in one species, but the numbers may
be much less in other species. Egg size may be con-
stant regardless of the size of the female, 1.e., those
laid in small numbers by minute species of Oligem-
bia appear to be as large as those laid by relatively
huge Antipaluria females.

The tightly clumped eggs are slightly slanted with
the opercula exposed. Many species spin a dense
covering of silk over the mass. Obviously, such cov-
erings, and parental guarding, reduce the percentage
of eggs parasitized by scelionid wasps. It is likely
that the habit of covering eggs is related to the geo-
graphic occurrence of the wasps. For example, in
the Mediterranean region where such wasps appar-
ently do not occur, the eggs of Haploembia spp. and
Embia spp. are uncovered and loosely clumped.

Species within unrelated Amazonian genera en-
close eggs in a sawdust-like matrix of chewed habitat
particles which is densely covered with silk, thus
forming a low mound on which the female rests (Fig.
42), ready to challenge approaching parasites and

FIGURE 18. As females guard their eggs, they lunge toward enemies—particularly egg para-
sites. Dinembia sp. (Embiidae). Northern Zambia.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 15

predators. However, such protection, like most de-
fenses in nature, is imperfect. For example, I found
within a mass containing 51 eggs, 12 fully developed
scelionid wasps clearly visible through transparent
egg shells.

Additional information on maternal protection of
eggs and young is provided by Edgerly (1987a, b;
1988, 1994) in her detailed study of plesiomorphic
Antipaluria urichi (Saussure) (Clothodidae) in
Trinidad.

FIGURE 19. Most embuids, such as Antipaluria Enderlein
(Clothodidae), reduce oviposition of wasp egg parasites
by packing a paste of pulverized material around their eggs.
Venezuela.

FIGURE 20. As eggs are laid on the sides of culture jars,
their number and imbedding sequence can be observed.
“Embia” surcoufi Navas, Mozambique.

Development

Adult female embiids exhibit little change in ap-
pearance from first instar nymphs except, of course,
for increased size and coloration. Ventrally, the eighth
and ninth abdominal paragenital sternites adjacent to
the vulva’s opening are modified, as is, of course,
maturation of internal reproductive organs. Neoten-
ic apterous males usually are similarly nymphoid but,
as adults, are more melanic, or pigmented, and have
distinct cranial and abdominal terminalia characters.

Males destined to have wings show the first ex-
ternal evidences during an early nymphal instar. At
first they are merely very slight extensions of the pos-
terior angles of the meso- and metascuta (Fig. 21A).
These are accompanied by increased development of
certain setae near, and on, the lateral margins of the
nota. The enlarged angles somewhat increase in size
during the stadium.

in
LDS

\\
‘\
\\
{

FIGURE 21. Wing development of a typical male embiid.
Oligotoma nigra Hagen.

In the next instar definite wing pads appear (Fig.
21B). Those of the mesothorax nearly reach the an-
terior margin of the metanotum. The lateral notal
setae now have increased in number and mark cours-
es of future wing veins. Development of the special
radius blood sinus vein (RBS) 1s indicated in some
specimens by faint “fleshy,” reddish lines which bor-
der RBS in the adult wing. As in the previous instar,
there is pad enlargement during the stadium.

16 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

In the next (penultimate) instar (fifth?) (Figs. 13,
21C, 22), the wing pads are well developed and great-
ly elongated; those of the mesothorax overlap most
of the metascutum, the pads of which extend caudal-
ly over most of the second abdominal scutum. They
are broadly attached to their respective scuta. How-
ever, these lines of attachment do not represent de-
finitive wing bases for they are actually part of the
posterior wing margin.

Definitive veins are indicated by their setae and
RBS by even greater pigmentation of its marginal
bands. Tracheae follow the same courses as the set-
ae. It is significant that venation of the pads con-
forms to that of the adult; MA being unbranched, for
example, in all oligotomid and anisembiid wings. The
observation by Melander (1903) that MA (his R4+5)
in Anisembia texana, as evidenced by the trache-
ation in the pads, is forked in the nymph and not in

the adult was probably an error, or anomalous.

Expansion of wings

During most of the penultimate nymph stage of
males destined to be fully alate, the wing pads are
clear, thin, and flat. Later they become thick, opaque,
cream white in color, and the dark lines bordering the
radial blood sinuses are conspicuous (Fig. 22). Inci-

dentally, at this time divisions of the tenth abdominal
tergite are visible through the derm and the adult cer-
ci are withdrawing basad. Finally, the nymph ceases
movement and in a short time it 1s possible to observe
ecdysis, and wing expansion. The following is my
account of these events in a species of Embiidae from
Africa’s Ruwenzori Mountains.

2:40 PM Nymph emerged from its last nymphal
exuvium. The unexpanded wings at first were slen-
der, strongly convex, thick, and cream white in color.

2:50 PM Starting from the costal and basal mar-
gins, the wings began to flatten and expand. Period-
ically, the soft adult wriggled and rotated its body.
The abdominal terminalia were distended, all struc-
tures were swollen, and the cerci projected laterad at
45°.

3:00 PM Entire basal half of the wings now broad-
ened and flattened. The distal half remained as small
and as narrow and fleshy as at 2:40 PM. At 3:10 PM
the entire costal margin had expanded with only the
apical end of the hind margin remaining fleshy. This
condition prevailed until 3:30 PM.

3:50 PM Left pair of wings now completely ex-

= =

®

FIGURE 22. Late penultimate instar of a male. Note that the wing pads have thickened and that cerci
of the adult are withdrawing from the nymphal skin. Antipaluria caribbeana Ross.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 17

panded, right pair remain fleshy ventrally at apices.
Wing veins paler than membranes.

4:10 PM All wings completely expanded, defin-
itive in form and thickness. Hyaline stripes had ap-
peared.

9:40 PM Wings now gray in tone. Body and leg
pigment developing.

7:30 AM (next day) Wings now smoke black.
Exuviae eaten during the night.

1:00 PM Male remains in same place in gallery.
Much darker in color.

9:00 AM (third day) Male still in same position.
Darker in color.

The above appears to typify wing expansion of
all Embudina. Sequences have been observed on
several occasions in distantly related species. How-
ever, a number of variations were noted, as follows:
In an embiid from Takoradi, Ghana, a male nymph
about to transform to an adult, was isolated at 2 PM.
As late as 8 PM it continued to spin silk even though
subdermal adult structures were well advanced and
in spite of the fact that it was destined to shed its
nymphal skin within two hours. During ecdysis the
head bowed downward and under towards the pros-
ternum until the abdomen completely withdrew from
the exuviae. The apex of the abdomen thrashed from
side to side to shake off the pelt. The unexpanded
wings were held out and away from the thorax and
bent caudad parallel to, but not touching, the sides of
the thorax. By 10:20 PM the wings reached the fourth
segment. The expanded bases were cream white while
the unexpanded apices remained smoke black.

In a new species of Dactylocerca from Alamos,
Mexico, an adult emerged at 10:40 PM and by 11:15
PM its wings had already assumed their definitive
shape. They were pure white and no veins or pattern
were evident. The exuviae was not yet consumed.
Surprisingly, while still teneral, the adult was able to
resume its silk spinning. By 8:00 AM the next day
the male was semi-hardened but the wings were still
pale. By this time the exuviae was almost entirely
ingested. At 8:00 AM, the third day, the male was
still in the same place, fully pigmented, and the exu-
viae had been completely consumed.

Parthenogenesis

Parthenogenesis probably occurs sporadically
throughout the order but only a few cases, involving

Mediterranean and equatorial African species, have
been studied. One of these, Parthenembia reclusa
Ross (1961), is widespread in western Africa. More
recently, I have decided that a number of undescribed
bisexual species from Angola and southeastern Zaire
must be assigned to Parthenembia, therefore, this
generic name is inappropriate. Scelembia virgo Ross
(1960) from Angola and Zaire also is parthenogenet-
ic. Other species of this potentially large genus are
bisexual. Caryiology of the above two species was
investigated by Renzo Stefani of Sardinia (1961).

Parthenogenesis of Haploembia solieri (Rambur),
indigenous to the Mediterranean region, was inten-
sively researched by Stefani and reported in a series
of papers cited in his bibliography. Rosanna Gior-
dano of the University of Vermont, who is fluent in
Italian, kindly provided the basis for the following
summary of Stefani’s conclusions on the bisexual and
parthenogenetic forms of H. solieri:

Stefani noted that bisexual H. solieri has only 20
chromosomes whereas the parthenogenetic form has
22. He assumed that the sex chromosome had been
replicated.

Haploembia solieri, like many other insects, suf-
fers infection by the gregarine sporozoan parasite,
Diplocystis clerci. Such infections aren’t well toler-
ated in bisexual populations of H. solieri for, although
both sexes are debilitated, the influence on males is
critical for they become ineffective mates due to dam-
aged sperm and lowered vitality. In contrast, a mi-
nority of females in a population which can repro-
duce parthenogenetically, tolerate infection and, even
though they produce fewer eggs, they can at least re-
produce. Asa result, parthenogenetic females sup-
plant sexual populations, as has happened on Sardin-
ia and Corsica.

During the 1960s when Stefani conducted his re-
search, small islands, such as Elba, had both sexual
and parthenogenetic individuals. He predicted that
with time only parthenogenetic populations would
exist on these islands. Perhaps by now this transfor-
mation has occurred.

Because the parthenogenetic form tolerates infec-
tion, and thereby permits D. clerci to mature and com-
plete its life cycle, H. solieri can serve as a vector for
the parasite.

Another of Stefani’s interesting observations was
that virgin females of the bisexual form of H. solieri,
unlike those of two other species of Haploembia he

18 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

checked, have the habit of eating their unfertilized
eggs. Very few eggs escape this cannibalism.

Strangely, 22-chromosome females (parthenoge-
netic), unlike 20-chromosome females (sexual), do
not eat their own unfertilized eggs. Sometimes para-
sites alter behavior of their hosts. However, infected
parthenogenetic females aren’t thus influenced.
Stefani was inclined to conclude that this results from
a duplication of the sex chromosomes, or perhaps a
combination of both—an interesting question to in-
vestigate.

Obviously, Stefani wished to determine how par-
thenogenetic females could have arisen from sexual
females. As an experiment he deliberately infected a
sexual female (20 chromosomes) with Diplocystis.
It managed to lay an egg in 1959, but this didn’t hatch
until 1961! The egg produced an adult female that
began to oviposit without mating. Much to his sur-
prise her progeny had a chromosome number of 22.
In effect, he had artificially produced parthenogene-
sis. He must have been amazed! [End of Giordano’s
abstract, somewhat reworded by me. ]

Because every individual is reproductive, the par-
thenogenetic form of Haploembia solieri readily be-
comes a “weed”; one rapidly spread and established
in new lands by human commerce. From California
where it was perhaps introduced in early Spanish
“Hides and Tallow” commerce, including dumping
ashore of sailing ship ballast, H. solieri is now a very
widespread species in warmer habitats of southwest-
ern United States and northwestern Mexico. I have
also found the parthenogenetic form common in many
Mediterranean mainland localities. Bisexual H. so-
lieri, and related species, or races, also occur, but these
apparently have not yet become parthenogenetic as a
result of Diplocystis clerci infection.

Much to my surprise, a bisexual population of H.
solieri was found in a garden (isolated from natural
environments) south of San Francisco, California. I
assumed that it resulted from a recent introduction in
nursery stock, perhaps from Spain. | attempted to
mate males with the very common parthenogenetic
females well established in nearby hills. The culture
produced only parthenogenetic broods. I didn’t ob-
serve copulation but, even if this had occurred, it is
likely that the parthenogenetic females were already
“self-fertilized.” This crossing attempt was made in
1976. It is possible that the sexual population has since
been eliminated by infection with D. clerci.

Diet

In nature embiids primarily eat weathered outer
bark of trees and decomposing leaf litter. They may
also eat mosses and lichens growing on bark, rocks,
termite mounds and soil surfaces. Undoubtedly, many
old substrates are coated or permeated with live mi-
croorganisms, such as algae, which are also nutritious.
There is no evidence that digestion is dependent on
symbiotic intestinal organisms.

It is likely that the diet of embiids is primordial
and that during the entire evolutionary history of the
order there has always been a certainty of food wher-
ever the insects choose to live on the basis of other
environmental factors, such as availability of crevice
retreats. Trees with exfoliating bark flakes, or verti-
cal crevices, are most likely to have embiid colonies
on them even though the nutritive value of the outer
bark of other tree species in the environment might
be the same.

Embiids seem to have no host plant preferences,
but one may expect that freshly-fallen leaf litter of
plants with strong antiherbivore chemicals, as in As-
clepias, Euphorbia, and Eucalyptus, will be avoid-
ed. However, such litter can be assimilated if
decomposing. For example, species of Australembi-
idae feed almost exclusively on a diet of layered, ag-
ing Eucalyptus leaf litter which also serves as the
habitat (Fig. 23). In Antipaluria intermedia (Davis)
of Venezuela, the dry season may be spent in leaf
litter and the wet season in sheet-like colonies on the
bark of adjacent trees.

A highly neotenic new genus and species from
the desert steppes of western Afghanistan extends
foraging galleries upward from subterranean retreats
into Artemesia shrubs to reach live foliage. In this
case, the small, chemically-protected, aromatic leaves
are transported into deep subterranean galleries as a
food supply during periods when the surface envi-
ronment is intolerably hot and dry. It can be assumed
that similar provisioning occurs in other species of
embuids inhabiting arid environments.

The universal acceptance of any non-toxic dead
leaves as food is demonstrated by my success in cul-
turing hundreds of usually-unrelated species from all
regions of the world on the same diet—dead Califor-
nia live oak leaves and a supplement of fresh Ro-
maine lettuce. Although not an essential element in

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 19

the diet, frequent replenishment of lettuce to the sur-
face of a laboratory culture increases carrying capac-
ity ina limited culture space while also moisturizing
the diet.

Adult females simply continue the diet of nymphs
of both sexes. Adult males, however—at least of all
non-neotenic species—never feed. This conclusion
is based on examination of thousands of adult male
specimens of hundreds of species during KOH mac-
eration while making microscope slides. These ob-
servations show that the gut of non-neotenic males is
invariably empty except during the short, teneral pe-
riod when it only contains fragments of its own pen-
ultimate nymphal exuviae which it ingested shortly
after the final ecdysis. This pelt gradually moves
caudad in the otherwise empty gut and eventually is
voided. Earlier workers, discovering such fragments
in the gut, erroneously concluded that embiid males
are predaceous. It is likely that all embuids, as do
many other insects with chewing mouthparts, invari-
ably ingest their exuviae after each moult. Some phys-
iologists believe that this insures a beneficial recycling
of sugars and nitrogen for chiton is a nitrogen-con-
taining polysaccharide.

Apparently, males of almost all species cease nor-
mal feeding during the penultimate instar and com-
pletely empty the gut prior to the final moult. This,
together with a reduced, or arrested, accumulation of
fat, results ina lighter, more vagile organism, but one
with shorter life expectancy

Neotenic, apterous, adult males of certain species
may, or may not, have the intestinal tract filled with
food in the usual stages of assimilation. Ingested
exuviae would be less discernible in such accumula-
tions. Males which continue to eat as adults are found
in several new genera in central Africa and in a race
of Metoligotoma reducta Davis of Queensland. Such
males, having a more pronounced intraspecific de-
gree of neoteny, have mandibles similar or identical
to those of nymphs and females and thus they are
suited for chewing food and apparently not for grip-
ping a female’s head during copulation.

Perhaps due to genetically-fixed, behavioral traits,
there is usually a correlation of taxon and specific
habitat and its particular food resources. For exam-
ple, in any suitable Queensland locality one encoun-
ters at least one species of each of the three families
occurring in Australia. Notoligotomids will be found
on rock, ledge, and bark surfaces without gallery ex-

FIGURE 23. Decomposing leaf litter of Eucalyptus is the
typical habitat and food of the numerous species and races
of Australia’s family Australembiidae. Such galleries ap-
parently don’t extend deeply into soil beneath the litter.
Metoligotoma ingens Davis. Black Mtn., Canberra, Aus-
tralia.

tensions into soil; australembiids will be encountered
between layers of dead leaves (usually Eucalyptus)
and, likewise, never extend galleries into soil; and
oligotomids, having a more widespread Australian
distribution, will be the only species dependent on
soil retreats. However, in the laboratory, species of
all three families thrive under identical cultural con-
ditions and eat the same food. Comparable correla-
tions of taxa and habitat may exist in any environment
inhabited by embiids. Therefore, an experienced col-
lector routinely examines each characteristic micro-
habitat as a means of securing cultures of all species
in the region.

It is probable that embiids seldom face food or
habitat limitations and this accounts in part for the
absence of striking biological and anatomical diver-
sity within the order. It is apparent that embiids are
never able to fully exploit local environments, or to
spread out into all suitable habitats. Typically, oc-
currence in any environment is spotty and many an
apparently satisfactory habitat, or even a major re-
gion, seemingly lacks representation of the order.

Such apparent absence may be attributed to haz-
ardous dispersal. Flight, because it is limited to males,
cannot extend a species’ range and the almost com-
pletely gallery-dependent, apterous females and
nymphs can walk only short distances outside of the
parent colony before encountering a predator—most

20 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

likely an ant. Unlike nuptial termites, ants, mayflies,
etc., embiids are unable to overwhelm the predation-
potential of an environment by concurrent bursts of
thousands of dispersing nuptual individuals in a lim-
ited time and space.

Another factor is the lack of a need to disperse
because of exhaustion of the food supply, or crowd-
ing. This is especially apparent in tropical evergreen
forests. For example, because food in the form of
weathered bark and surface growths would be restored
very soon after consumption, it is possible that suc-
cessive generations of one or more species of embi-
ids could remain indefinitely on a single, large tree
trunk. The activities of parasitoids, diseases, and other
natural hazards would also tend to limit embiid pop-
ulations.

Because of reduction in dispersal incentives and
limited vagility, embiids promise to serve as excel-
lent indicators of zoogeographic regions and conti-

nental drift.

Movement

Embiids are especially adapted for movement in
narrow galleries. Shortness of legs, probably a plesi-
omorphic feature retained from a Paleozoic arche-
type, is especially important. Such ancestors probably
depended on bark, rock, plant crevices, or layered
leaf litter, as refuges from predators. Because such
retreats usually are edible, embiids didn’t have to
venture far afoot, or in flight, to reach food. Howev-
er, even short forays to extend “grazing” were made
safer by evolution of an ability to produce silk cover-
ways.

Although wings were once possessed by adults
of both sexes, rapid flight probably never was impor-
tant as a means of avoiding predators. However, aerial
dispersal of alate females of ancestral species must
have fostered spread of taxa over a long period of
geological time, thus extending the order’s range
throughout warm portions of Panagea. Later, how-
ever, except for sporadic movement of gallery sub-
strate objects by storms, bird plumage, or in
commerce, dispersal is severly limited by universal
female apterism. Therefore, walking by females is
the only way an embiid can move to a new location if
the one presently occupied becomes intolerably wet,
dry, exposed, or “overgrazed.”

Ordinarily, nymphs remain within their galleries
unless they are torn open by predators, such as birds,

mammals, or army ants. Adults of both sexes, how-
ever, may concurrently leave their galleries soon af-
ter the first rains following a long dry season, or during
what appears to be primordial, prenuptual excitement
comparable to that causing simultaneous nest-exo-
dus of nuptual termites and ants. However, because
embiid colonies are less populous and scattered, nup-
tual embiids never create noticeable swarms. Dis-
persal of individual embiids are therefore probably
more vulnerable to predators. Consequently, some
diurnal species of embiids have evolved a degree of
protection by mimicking appearance and movement
of unpalatable or dangerous models, such as ponerine
ants, paederine Staphylinidae, or lycid or pyrochroid
beetles.

Forward walking, either within or outside of the
galleries, is steady, slightly sinous, with all legs in-
volved. Stimulation of the cerci triggers rapid for-
ward bursts of speed. However, the most important
defensive movement is rapidly backward, powered
by the enlarged tibial flexor (depressor) muscles that
almost fill the large hind femora. A firm tread, espe-
cially ona silk surface, is insured by numerous, stout
setae on the plantar surface of the hind basitars1.
Backward movement has had a profound influence
on wing specializations, as well as being the primary
cause of neotenic apterism and subapterism.

During casual forward movement outside of gal-
leries, adult females often walk with their genital seg-
ments arched upwards as though to “welcome”
insertion of a male’s terminalia. Adult males of many
species often walk with the abdominal apex bent for-
ward on the dorsum of the abdomen. In alate species
this apex may even press tightly against a correspond-
ing forward fold of wing apices and must be an addi-
tional means of reducing the adverse barb effect of
both wings and terminalia during reverse movement
within galleries. Such forward bending of the ab-
dominal apex also occurs in apterous males of some
species.

Males outside of galleries are very lively and alert.
Often they stand high on the forelegs, thus elevating
the usually large-eyed head, and the prothorax. A
male’s head is very mobile and capable of turning at
least 45° from the longitudinal axis. Especially while
resting on walls beneath artificial lights, males often
twist the head in a mantid-like fashion as they follow
the movements of an observer. Just before flight, a
male’s body may tremble, antennae vibrate and twirl,

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 21

and the forebody bobs up and down (Fig. 11). Alate
males thus retain much of the activity and sensitivity
of free-living ancestors while the nymphs and adult
females, perhaps because they are secluded in
galleries, are less alert.

Embiids are highly sensitive to vibration, as may
be observed in laboratory cultures kept in jars. In
these, adults often rest for long periods in upper gal-
lery levels. In reaction to human approach, or vibra-
tion, even as much as ten feet away, embiids suddenly,
often in unison, back downward into the depths of
the culture. Because of this, a person wishing to
collect such individuals, must approach a culture jar
slowly, gently open its lid, and trap the desired indi-
vidual by blocking the gallery behind it. At times
embiids feign death, even during handling, and then
suddenly burst into activity.

Habitats

1. TROPICAL EVERGREEN FORESTS.
Such forests appear to be the basic, or primordial en-
vironment of the order. Most species occurring in
wet forests are arboreal, or colonize sheltered sur-
faces, such as undersides of ledges, logs or branches
which remain relatively dry during frequent rains.
Even on well-drained, vertical tree trunks, or road
banks, the sheltered slant-side is preferred (Fig. 24).

In spite of the great number of species potentially
present in a particular forest with its profusion of
microhabitats and food, an experienced collector may,
after hours, even days of concentrated search 1n a vir-
gin forest, fail to find a single colony. This difficulty
apparently results from the abundance of ant preda-
tors, diffusion, and an inability of embiids to fully
exploit or reach potential habitats.

An entomologist, regardless of his specialty, is
well aware that the best general insect collecting is
found in recently disturbed forests and this is partic-
ularly the case with embiids. One soon learns that
embiid colonies are most frequently encountered on
residual trees, stumps and logs left in forest lands
partially cleared (but not burned) for plantations, such
as those of cocoa and coffee. Also, clearance and
trails provide easier collector access to potential sites.

Natural and artificial habitats within tropical for-
ests are listed as follows:

(a) Surfaces of trunks, limbs and lianas. Some
embiid species spin conspicuous galleries of clean
silk fully exposed to view (Figs. 25, 26). As weath-

ered outer bark is consumed, galleries are extended
over fresh surfaces. Occasionally, an entire tree trunk
is matted with the silk of apparently merged galleries
of separate broods (Fig. 26). It is conceivable that
regrowth of an edible substrate is delayed by such
cover and that portions of a trunk will have to be aban-
doned. While large colonies are active, the matting
of silk constitutes a protective cover, which, as stated
before, is almost as effective as a layer of bark for it
not only protects new galleries, but also other organ-
isms taking advantage of the cover.

Often, for some distance around a large colony,
small satellite colonies will be found on adjacent trees.
Farther out, the forest may lack additional colonies
until another concentration occurs. In some species
gallery silk on bark is made inconspicuous by a cov-
ering of pulverized fecal pellets and bark fragments
deliberately placed on the surface by the embiids
(Figs. 8,9). Often minute plant life will grow on this
and render the silk even more obscure while also en-
hancing the protective cover.

(b) Bark crevices and flakes. At times galleries
of small species are not visible without removal of
bark flakes. However, presence of a colony is often
indicated by a slight extension of silk beyond the edge
of a flake.

(c) Roots of orchids and other epiphytes. Vines
and their leaves appressed to tree trunks often shelter
colonies. For example, the flat, circular leaves of
Peperomia rooted to tree bark provide excellent cov-
er for colonies of Saussurembia Davis in Costa Rica.

(d) Undersides of elevated logs often criss-
crossed in clearing and selectively-logged forests.
Recently-felled trees provide access to colonies nor-
mally out of reach on standing trees. Because of
changed exposure, the colonies may shift to the un-
dersides of levelled trunks and branches, thereby gain-
ing greater shelter than that afforded on upright trees.

(e) Fence and stockade posts. Especially those
recently cut which still have loose bark, are frequent
sites of colonies.

(f) Surfaces of ledges and earthen banks. Es-
pecially if somewhat under-slanted and favorably ex-
posed toward or away from the sun, these surfaces
tend to serve much like tree bark in that they offer
many retreat crevices and a food supply in the form
of surface growths.

(g) Surfaces and crevices of structures. Even
mossy, Steel girders of bridges.

22 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FIGURE 24. Typical rainforest habitat (semicleared). Several embiid species may be
found on such a trunk. Near Belém, Brazil.

2. TROPICAL CLOUD FORESTS. These oc-
cur at various altitudes, usually starting at 3,000
meters on wet, often windward slopes of tropical
peaks and ranges. Embuids reach their highest known
altitude in this zone—about 3,500 meters in the Ec-
uadorian Andes. In spite of cold nights and proxim-
ity to snow, each day is warm in equatorial latitudes.
Therefore, there is no need for the special, low tem-
perature, physiological adaptations, required in tem-
perate regions with prolonged cold periods.

Males of many cloud forest species are slender-
bodied, have disproportionately long legs and large
wings (Fig. 12). All are difficult to culture and it is
advisable to secure adequate series of adults in the
field by persistent collecting rather than to depend
on culturing. Late instar nymphs may survive long
enough to mature, however.

All habitats listed for lowland rain forests should
be searched for embiids. However, colonies are not
only less conspicuous, but also less accessible, be-
cause of steep terrain, usually impenetrable vegeta-
tion (especially bamboo), and dense coverings of

mosses and epiphytes on most surfaces. Because of
such disadvantages, the best collecting opportunities
are in agricultural clearings and in edges of residual
gallery forest.

The most characteristic cloud forest embiid hab-
itat is moss festooning from tree limbs and trail banks.
Such moss is often thoroughly bound together with
embiid galleries (Fig. 27).

Because of better drainage, shelter, and accessi-
bility, road and trail banks are the best places to search
for embiids in cloud forests. At times galleries on
road banks can even be seen from a moving vehicle.

3. SEASONALLY-DRY GRASSY WOOD-
LAND. In the rainy season these forests are lush and
green, but the long dry season and repeated fires tend
to reduce epiphytic growth and thus suitability of most
tree trunk surfaces as embiid habitats. However, this
zone—especially the savanna woodlands of central
Africa and the deciduous woodlands of India and
southeast Asia (Fig. 28)—is home to some of the most
interesting species of the order. It should be noted
that evergreen groundwater forests frequently occur

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 23

FIGURE 25. In rainforests, colonies often are conspicuous.
It is surprising that they are so often neglected by ento-
mologists. Clothoda longicauda Ross. Tingo Maria, Peru.

within this type of woodland along rivers and on shad-
ed slopes. The characteristic dry woodland habitats
are:

(a) Tree bark crevices and flakes. Several large,
pale species are found in bark, but tend to develop in
individual galleries, nymphs having dispersed soon
after hatching. At certain seasons only tiny nymphs
of potentially-large-bodied species will be found only
after careful search. Species of African genera, Ber-
landembia Davis and Dinembia Davis, are exam-
ples. Occasionally, seemingly-empty, prior-season
galleries will be found, but close examination will
reveal unattended live egg masses or tiny nymphs
enduring the dry season, above the height of bush
fires.

(b) Dead branches attached to trees. These are
especially important habitats. In them, man-made
splinter-cracks, beetle burrows, pithy twigs and loose
bark serve as refuges from predators, temperature
extremes, and dessication.

FIGURE 26. Interconnected galleries of many broods of
Machadoembia n.sp. almost cover buttresses of huge trees
in Congo rainforests. Such concentrations occasionally are
ripped open and plundered by vertebrate predators.

(c) Dead branches and logs on ground. Espe-
cially those recently cut or fallen should be rolled
over and their undersides carefully examined. Gal-
leries once connected with those in thus-exposed leaf
litter and soil crevices should also be searched.

(d) Leaf litter on forest floor. Embiid galleries
from deep, protective, soil crevices may extend into
such cover during the wet season, or cooler periods
of the day or night. Species of the Australian family
Australembiidae almost exclusively occur amongst
matted leaves and apparently do not utilize soil re-
treats (Fig. 23).

(e) Crevices in surfaces of rocks, ledges and
termite mounds. The latter offer almost rock-hard
protective retreats, good drainage in the wet season,
and the weathered surfaces are often rich in nutrients
(Figs. 5, 6).

(f.) Leaf clusters in understory of savanna wood-
land (Figs. 28; 29A, B).

24 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FIGURE 27. In cloud forests many unrelated species utilize
well-drained, festooned moss both as habitat and food. New
genus and species (Oligotomidae). Gunong Batu
Brinchong, Malaya.

4. SEMI-ARID, OPEN GRASSLANDS. When-
ever present, all previously mentioned habitats should
be searched in this zone. Forest habitats may be
localized along river courses and around springs.

(a) Stones. Most grassland species are best en-
countered under stones (Figs. 30, 31) but it should be
realized that the turned stone simply exposes a soil
profile (Fig. 32) for it is likely that most grassland
embiids are widely distributed in the sod and are not
necessarily stone-cover-dependent. Very little sur-
face activity will be seen in the dry season and exca-
vations may be required to secure cultures (Fig. 33).
However, there is a chance of encountering embiids
in surface galleries at night or early in the morning.
Their attraction to the surface could be increased by
artificially wetting soil around galleries expected to
contain embiids.

(b) Dead branches and limbs riddled with bee-
tle borings. Some embiids utilize these retreats to
survive the long dry season and fires.

(c) Leaf litter and soil. At the bases of large trees
and beneath clumped shrubs.

(d) Soil crevices in open ground. \n western
Australia, immediately after rains, species of
Aposthonia extend galleries in soil crevices upward

FIGURE 28. The savanna (Brachystegia) woodlands of the Congo-Zambezi divide have
the greatest diversity of higher embiid taxa. “Nests” of female Enveja (Fig. 29A) were
present on low shrubs in this scene.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 25

FIGURE 29A. Silk webs in a leaf cluster. Such leaves may
be dead or alive on low understory vegetation. This habitat
is chiefly used by species of Enveja Navas occurring in cer-
tain portions of Africa’s Brachystegia woodlands. By oc-
cupation of leaf clusters, females and their early brood es-
cape excessive soil moisture in the rainy season. With rain-
fall decline, the embiids move down into leaf litter and, fi-
nally, into soil cracks to escape dry conditions and fires.

He i we
2% RDG Speeder em a

FIGURE 29B. Opened Enveja nest revealing female guard-
ing eggs. Unlike those of most embiids, the eggs are loosely
covered with fibrous debris—perhaps her fragmented fe-
cal pellets.

to reach leaf litter food. Fragments may be carried
down into subterranean galleries to serve as food pro-
visions. With a return to complete aridity, temporary
surface galleries may soon weather away and thus no
evidence of embiid occurrence may remain. An ex-
ceptional species in western Afghanistan extends gal-
leries upward into low shrubs (Artemesia) to collect
leaf fragments.

5. DESERT AREAS. An embiid fauna is consid-
erably reduced by extreme aridity and may be con-
fined to oases, drainage lines and foggy coastal
deserts. A lack of significant precipitation almost
completely eliminates occupation of all above-ground
habitats. However, palm trees, such as the date palm,
provide safe retreat in leaf bases of the trunk which
tend to collect abundant embiid food in the form of
leaf debris. In the Nile Delta two species of the or-
der, Embia savignyi Westood, and Haploembia solieri
(Rambur), occur in nest material deep in the under-
ground burrows of rats of the genus Arvicanthus
(Hoogstraal, pers. com.) (Figs. 34-36).

6. HUMAN HABITATS. Cities may have large
embiid populations on the bark of shade trees. Most
often these are introduced species of the India-cen-
tered genus Oligotoma: O. saundersii (Westwood),
O. humbertiana (Saussure), and O. nigra (Hagen).
Native species may also be found, especially if the
highly disturbed areas are near natural environments.
In Trinidad extensive webs of a native species,
Antipaluria urichi (Saussure), are conspicuous on
trees along streets and in parks.

Occasionally, embiids eat stored products. For
example, species have been reported in sugar refin-
eries in Australia.

In Perth, Australia, Notoligotoma hardyi (Frieder-
ichs) inhabits old wood fences in residential areas uti-
lizing crevices as retreats and weathered wood
surfaces as food. In southern California galleries of
Oligotoma nigra Hagen occasionally extend from the
ground up the sides of residential foundations, but
embiids do no damage to structures.

In vineyards embiids may extend their webs into
bunches of grapes near the ground, but probably feed
only on dead material and detritus accumulated in
interstices, not the grapes. In Senegal, Oligotoma
saundersii (Westwood) may produce colonies in piles
of harvested peanuts accumulated prior to export.

26 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FIGURE 30. An excellent stony, seasonally-dry embiid habitat in southern India. (David
Cavagnaro, assistant during Indo-Australian expedition).

FIGURE 31. This dry slope in central Algeria, too stony for FIGURE 32. A removed stone exposes a sod profile with
5 g J I

plowing, is the type locality of Embia silvestrii Davis galleries of Haploembia solieri (Rambur) (Oligotomidae)

(Embiidae) in Californian seasonally-dry grasslands. From cool moist

depths, embiids move upward to feed on dead plant litter,
perhaps at night or other cool periods.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 27

FiGure 33. In arid Baja California, Mexico, Bulbocerca
sini (Chamberlin) (Anisembiidae), feed at night and dur-
ing the rainy season, on leaf litter accumulated between
desert stones. Dense silk galleries extend downward in the
silty soil.

FIGURE 35. Deserts near sea coasts may have a special

embiid fauna supported by fog precipitation. For example,
in the almost rainless coastal fog zones of Peru and south-
western Angola, embiids may extensively colonize fog-sup-
ported lichens growing on rock and ground surfaces. Re-
treats beneath rocks and into soil and rock crevices pro-
vide escape from the excessive dryness of certain seasons
and/or hot periods of the day. Illustrated is the habitat of
Chelicerca n. sp. (Anisembiidae) in the Peruvian Desert
north of Callao.

FIGURE 34. Numerous galleries of Notoligotoma n. sp.
(Notoligotomidae) are found, but not exclusively, on
undersurfaces of exfoliating slabs on “granite islands” in
southwestern Australia. Wind-blown plant debris caught
beneath such slabs is the basis of an ecosystem which in-
cludes many embiid colonies.

FIGURE 36. Galleries of some species occur on the sur-

faces of roots of perennial shrubs as well as of trees. An
interesting new genus and species was thus found in A frica’s
Namib Desert following the roots of Welwitschia plants.
However, the species probably isn’t restricted to a host.

28 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

Geographic range

Embiids are warm-climate-adapted insects whose
natural occurrence is almost universal in all suitable
environments on any continent, or continental island,
which has a tropical or warm-temperate climate. The
order’s distribution roughly coincides, for example,
with that of phasmids and termites but the latter ex-
tend farther into colder latitudes, perhaps because of
deeper penetration into habitats, such as soil, logs,
and buildings. But, as with termites, embiid occur-
rence in temperate regions is reduced. In western
North America the most northerly record of an en-
demic species, Dactylocerca rubra (Ross), is in cen-
tral Utah, lat. 39°N. Non-endemic Haploembia solieri
(Rambur) occurs farther north into eastern Washing-
ton. In eastern North America, represented by Dira-
dius vandykei (Ross), the order reaches about 35° N
in coastal Virginia.

In South America the southern limit has not been
determined. The most southern occurrence known
to me is Pararhagadochir trachelia (Navas) in a
desert habitat west of Mendoza, Argentina, at about
35°S. West of the Andes, the southern range appears
to be limited by the extreme aridity of the Atacama
Desert of northern Chile, rather than by latitudinal
cold. An undescribed species of Chelicerca Ross
probably extends well southward in the coastal fog
belt (lomas) of northern Chile, but, to date, | have
found it only as far south as Mollendo, Peru. The
northern Chilean lomas aren’t very accessible to a
collector. There is absolutely no evidence of a South-
ern Hemisphere, or Patagonian, origin of any spe-
cies.

In the Old World, endemic species of the order
occur in warm portions of all continents and conti-
nental islands even including Tasmania [Mero-
ligotoma tasmanica Davis and Aposthonia gurneyi
(Froggatt)]. New Zealand and Madagascar, which
apparently could not be reached by natural spread of
any species, evidently do not have endemic repre-
sentation of the order. The northern extremity of range
in western Eurasia is in the Crimea, about 45°N, with
Haploembia solieri (Rambur). In the Middle East,
Parembia persica (McLachlan) ranges into northern
Iran, Afghanistan, and Turkestan (about 38°N). In
eastern Asia, Aposthonia japonica (Okajima) reach-
es about 32°N in southern Japan, but this may be a
species introduced from a south-Asian locality. It is
assumed that the order ranges to comparable latitudes
along the coast of mainland China.

Oceanic islands of the Pacific appear to have no
endemic species, those present having been intro-
duced by movements of ancient and modern man.
Aposthonia oceania (Ross) 1s widespread in Oceania
as far south as Easter Island 27°S and New Caledonia
(20°S ), at first by dispersal of Polynesians.

The fossil record indicates occurrences far to the
north of present ranges, but this may be due to drift
of land surfaces northward from warmer zones sub-
sequent to fossilization. In the case of the Florissant
fossil, Lithembia florissantensis (Cockerell), the fos-
sil beds probably drifted northwestward and were lift-
ed an additional 5,000 feet in altitude by elevation of
the Rocky Mountain Range. I speculated that Baltic
Amber fossil, Electroembia antiqua (Pictet), lived
when the Baltic land surface was in a warm, dry Med-
iterranean latitude (Ross, 1956).

Ecological range

Within the extremes of geographic range, embiids
are confined to regions lacking prolonged cold peri-
ods; the more equatorial the location, the greater the
potential altitudinal range. Aridity can also be a lim-
iting factor, but some practically rainless deserts, such
as the Peruvian and Namib, have one or more embiid
species in habitats regularly dampened by sea fog.

The primordial habitat of the order appears to be
tropical rain forests. Those of the Amazon Basin re-
tain the order’s most “primitive” species. From such
forest centers, embiids appear to have radiated and
become adapted to many other types of environments.
In this movement the trend has been toward an eva-
sion of adverse ecological conditions rather than phys-
iological adaptation to them. Thus, in regions
experiencing prolonged dry seasons, embiids escape
heat, dessication and fires by retreating downward in
silken galleries following cracks, which provide ac-
cess to cooler, more moist soil depths. Others may
find refuge in burrows produced by beetle larvae in
dead limbs of trees. Under such circumstances the
activities of the embiids are confined to the rainy sea-
son or to periods of the day or night when the surface
temperature and moisture are favorable.

In equatorial regions embiids spreading from op-
timum lowland zones have adapted to higher eleva-
tions. The highest known altitudinal record of the
order is that of anew genus and species found in cloud
forests and paramos above Cuenca, Ecuador at about
3,500 meters elevation. The order is present in most

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 29

of the cloud forests I have visited, e.g., Andean, cen-
tral African, and Malaysian. In such high zones the
low temperature extreme may not be much more se-
vere than that of many warm-temperate regions at sea
level. The zones are also subject to almost daily in-
tense equatorial, solar heat at all seasons. A few
endurable hours of cold occur one night at a time rath-
er than for weeks and months, as in cold latitudes.

Low temperature, however, appears to be the most
critical ecological factor limiting the order’s occur-
rence. Perhaps this is due to an absence of a truly
cold resisting or hibernating life stage. No species is
known to utilize the egg stage as a means of surviv-
ing cold periods, and, of course, there is no dormant
pupal stage. Three species endure rather severe win-
ter cold: Dactylocerca rubra (Ross) in central Utah,
Anisembia texana (Mel.) in northern Texas and south-
ern Oklahoma, and Haploembia solieri (Rambur) in
Crimea and, by introduction into eastern Washing-
ton. Haploembia solieri overwinters as half-devel-
oped nymphs which, during cold periods, are
protected in dense cocoon-like silk enclosures. Such
a protection probably is characteristic of all cold-en-
during species of the order. These “cocoons” proba-
bly are a means by which predation is reduced during
the embiid’s torpid state for, in cold environments,
potential predators, such as carabid beetles, are like-
ly to be adapted for movement during cold periods
and therefore can easily catch unprotected prey.

Natural dispersal

The major distribution of the order must have oc-
curred on portions of Pangaea during the long period
when both sexes possessed wings. After females be-
came universally apterous, natural dispersal outside
of galleries became slow and hazardous, especially
after ants evolved to become significant predators. It
is possible, however, that the natural spread of small
species, such as teratembiids, could have occurred as
substrate objects, such as twigs and branches, were
moved by wind, and rafting. On or in such objects
survival and transport of embiids would have been
insured by enclosure in securely attached silk galler-
ies. When continents, such as South America and
Africa, were still separated by a narrow sea, aerial
movement and rafting would have been more likely
considering the long time periods available.

Another, yet rare means of dispersal could be in
bird plumage, as is likely to have happened in the

case of Chelicerca galapagoensis Ross. Colonies of
this anisembiid frequently web the general environ-
ment and nests of birds in higher elevations in larger
islands of the Galapagos. It would seem to be a sim-
ple matter for embiids to extend their galleries into
the feathers of birds resting on a nest and, although
many of these would be groomed out of the plumage,
a certain percentage over the ages, could be carried
from place to place and thus experience an extension
of range. Chelicerca galapagoensis is related to a
Peruvian and Ecuadorian loma (fog zone) Chelicer-
ca and it is likely that a limited gene pool of a future
new species was carried to the islands from the main-
land as “passengers” in the plumage of birds (Ross,

1966).

Dispersal by man

Because of the use of crevice retreats and the ad-
herence of silk galleries to logs, plant materials, ship’s
ballast and cargo, the range of some species has been,
and will continue to be, extended when such objects
are moved either by natural or artificial means. In
warmer regions, especially where man has engaged
in commerce for thousands of years, it is probable
that species which appear to be endemic, were actu-
ally unintentionally introduced by humans. This is
particularly likely in southern Asia and the Mediter-
ranean region. It is to be expected that the greater
speed and volume of modern commerce will acceler-
ate artificial spread of additional species.

Embiidina are able, at least temporarily, to be-
come established in greenhouses located in temper-
ate regions. One of the earliest known species,
“Embia” michaeli McLachlan, 1877, was collected
in an orchid house in England and is known only from
its incomplete type specimen. The origin of the or-
chid appears to have been northeast India, or Burma.

The establishment of a South American teratem-
biid, apparently Diradius intricatus (Davis), in an
orchid greenhouse near Wageningen, Netherlands was
reported to me by R. H. Cobben. It was first noted
about 1970 and, in spite of several intensive insecti-
cidal treatments, the infestation persisted many years
until the greenhouse was removed for other reasons.
It is unlikely, however, that such an introduction would
have economic significance. Undoubtedly the insects
were unintentionally introduced on live plants, per-
haps orchids, from Suriname.

30 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

For many years embiids were frequently inter-
cepted by U. S. plant quarantine inspectors in orchid
plants from various Neotropical sources—particularly
Colombia. It is unlikely that any of these would have
survived very long, even if they slipped past inspec-
tion.

Some embiid species are more successful than
others in becoming established abroad and it is sig-
nificant that most of these belong to the family Olig-
otomidae. Certain species of this family often have a
high survival potential and reproductive vigor as in-
dicated by the fact that they are the easiest to propa-
gate in laboratory cultures. They also tend to produce
overlapping generations throughout the year.

In order of importance, species whose range is
steadily being extended by man are listed, as follows
(Species 1—9 are in Oligotomidae):

1. Oligotoma saundersii (Westwood).

Endemic to northern India. Now likely to be found
in any warm, moist tropical or warm-temperate re-
gion. This is the species most often collected in set-
tled areas by the non-specialist. Males frequently are
attracted to lights.

2. Oligotoma humbertiana (Saussure).

Endemic to India. Now very common in Indone-
sia, Philippines, China, western Mexico (probably in-
troduced from the Philippines in galleon trade to
Acapulco). Unfortunately, this species is becoming
distributed in natural areas of western Mexico, espe-
cially in the states of Sinaloa and Nayarit. Males fre-
quently appear at lights.

3. Oligotoma nigra Hagen.

Probably endemic to northern India, O. nigra has
spread westward and northwestward in ancient cara-
van traffic and now is well established in Arabia, the
Middle East and the valley of the Nile. It was appar-
ently introduced into southern United States in date
palm cuttings (Ross, 1957) and now extends into east-
ern Texas and northwestern Mexico. It has also been
collected in inland NSW Australia. It is now occu-
pying natural habitats. Males frequent at lights.

4. Oligotoma greeniana Enderlein.

Probably endemic to India. Now established in
Malaya, Taiwan, Mindanao, Hong Kong, Macao
(probably from Goa), and China’s southwestern in-
terior.

5. Aposthonia borneensis (Hagen).

Probably endemic to continental southeastern
Asia, or Indonesia. Now also occurs in southern
China, Vietnam, Philippines, New Guinea and Indo-
nesia.

6. Aposthonia oceania (Ross).

Possibly endemic to continental southeastern Asia,
or Indonesia. Apparently was spread by early Polyne-
sians to various portions of Micronesia and Polyne-
sia, including Easter Island and New Caledonia. In
at least one case, Aposthonia micronesiae (Ross),
from Mariana Island, speciation may have since oc-
curred but there is a possibility that this is a species
introduced from a source-region distinct from that of
A. oceania.

7. Aposthonia ceylonica (Enderlein).

Probably endemic to southern India and Sri Lan-
ka. Now spread to Mauritius, Madagascar, Malaya,
Thailand and Laos.

8. Aposthonia minuscula (Enderlein).

Probably endemic to India. Now spread to East
African coast and Madagascar.

9. Haploembia solieri (Rambut).

Endemic to some undetermined Mediterranean
area. Now common almost throughout Mediterranean
shores and Black Sea region. This wide range per-
haps resulted from ancient land and sea commerce.
A parthenogenetic form occurs in scattered Mediter-
ranean regions, especially on certain islands, and has
since been spread to southern Spain, northwestern
Africa, Asia Minor, Afghanistan, the Canary Islands
and western United States—especially California.

10. Embia savignyi Westwood.

Probably endemic to southern Sudan and adja-
cent regions. Now spread westward as far as north-
ern Nigeria and north into the Nile Delta, Israel, and
Crete. Males fly to light.

11. Parembia persica (McLachlan).

Probably endemic to northwestern India, or Pa-
kistan. Now spread, perhaps at first by ancient com-
merce, as far north as Russian Turkestan and as far
west as Israel, Jordan and Arabia. The damaged type

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 31

of Parembia producta Davis from Somalia appears
to represent P persica (Ross, 1981).

12. Pararhagadochir trinitatis (Saussure).

Probably endemic to northeastern South Ameri-
ca and Trinidad. Now sporadically common on tree
trunks around San Jose, Guatemala; northern Costa
Rica; Panama; and probably elsewhere.

13. Some South American species, such as
Archembia batesi (Mc.L.) and Pararhagadochir bic-
ingillata (Enderlein), appear to have extended their
range in river commerce.

Natural hazards

PREDATORS: Outside of their galleries embi-
ids are easy prey to predators, especially ants, as well
as spiders, beetles, centipedes, and small vertebrates.
Predation appears to be the prime factor limiting es-
tablishment of colonies in new places. Out of their
galleries it is difficult to think of creatures more vul-
nerable to predation than embiids. They cannot run
or fly very well, they lack biting or stinging defenses,
repugnant secretions, body armor, irritating vestiture,
or defense through high reproductive potential. There
is indication, however, that some diurnally-dispers-
ing species may reduce predation by mimicking the
coloration and movement of repugnant or dangerous
models. For example, in the Neotropical region many
species, especially females, resemble notoriously ag-
gressive, stinging ponerine ants. Most spectacular is
the orange-and-black pattern, especially of nymphs
and females, of several species of Dihybocercus
Enderlein in south-central Africa. Combined with a
similar body form, such embiids strongly resemble
aposematic, highly irritating staphylinid beetles of the
genus Paederus.

The most significant defense, although not com-
plete, is confinement in silk galleries. The impor-
tance of silk gallery protection was clearly evident
during a brief field experiment conducted by me in
Singapore with colonies of Oligotoma saundersti
(Westw.). The colonies occurred on bark of shade
trees on which swarmed large, vicious Oecophylla
ants. These ants ran over the silk webs apparently
unable to detect embiids underfoot. However, when
I opened some webs, the exposed embiids were im-
mediately detected and carried off by the ants. Un-
der such circumstances the spread of the embiids
would be possible only by gallery extension, or by

FIGURE 37. Muirid-like appearance of a species of
Plokiophilidae bug, Embiophila (Acladina) africana
Carayon, I collected in Katanga, central Africa. A. Male.
B. Brachypterous female. (After Carayon)

movement outside of galleries during inactive peri-
ods of predators. It is likely that ants are the most
significant predators of embiids and that the major
extension of embiid occurrence on Earth took place
before ants evolved and became abundant. Of course,
during the pre-ant period, wide dispersal was aug-
mented by presence of wings in both sexes.

In addition to arthropod predation, colonies may
be ripped open and plundered by birds, rodents, le-
murs, monkeys, and other vertebrates. This is most
evident in tropical forests where one occasionally
encounters sheet-like mats of galleries torn open and
devoid of occupants following discovery by an in-
sectivore.

As is to be expected in any ancient group of in-
sects, embiids provide both habitat and food for par-
asitoids and parasites, some exclusive to the order.
Included are various species of tiny, brown,
anthocorid-like Hemiptera of the little-known family
Plokiophilidae (Figs. 37, 38). All members of this
tropical family live on silk webs; those of the sub-
family Plokiophilinae appear to be confined to spi-
der webs and the Embiophilinae to galleries of Em-
biidina. Apparently, however, a substrate of silk and
available food is more important than actual associa-
tion with spider and embiid hosts for Carayon reared
the bugs for several generations on a diet of mites.
Carayon’s work (1974) is the best reference.

32 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

FIGURE 38. Plokiophilid bug nymph sucking fluids from
the basitarsus of a dead embiid (Archembia n. sp., Tingo
Maria, Peru).

I have encountered the bugs sporadically in colo-
nies of many species of embiids in both Old and New
World tropics. The family appears to be absent in
Australia. Often they are inadvertently collected as
silk galleries and other habitat material is included in
a field culture. As embiid cultures grow, embiophil-
ines usually thrive, but apparently do not always kill
their hosts. The bugs are most numerous in the vi-
cinity of embiid egg masses and broods of young
nymphs. Carayon (1974) reported that they regular-
ly suck fluids from eggs and young nymphs. Tyro-
glyphid mites are always present in embiid cultures
and must supplement the bug’s diet. I have also ob-
served them sucking shriveled, recently dead, embiids.
Unless salivary toxins are introduced in their feed-
ing, the bugs should not adversely affect the embiids
unless several individuals simultaneously feed on a
single egg or a small nymph.

The following arthropods are also encountered
within embiid galleries:

Acarina. Predatory mites frequently infest cul-
tures and can seriously reduce their vitality. Larval
mites of the family Trombiculidae attached to embiids

have been encountered. Scavenger mites are also
abundant in most colonies.

Rhaphidioptera. The sinuous, predaceous lar-
vae of snake flies frequently are encountered in un-
der-stone galleries of Haploembia solieri (Rambur)
in California.

Coleoptera. On one occasion (Embia sabulosa
group, S.W. Africa) I encountered numerous larvae
of a species of Lampyridae which had consumed all
of the embiids in an under-stone colony. I found a
beetle of the family Monommidae in galleries of
Clothoda longicauda Ross at Tingo Maria, Peru.
Myers (1928) reported numerous monommids,
Hyporrhagus marginatus (Fab.), within extensive,
mat-like webs of Mesembia hospes (Myers) in Cuba.

Diptera. Larva of Leptopteromyia Williston
(Leptogastridae) have been reared on several occa-
sions in embuid cultures from widespread Neotropi-
cal localities. In the labyrinth ofa culture it is difficult
to determine the exact relationship of these slender
fly larvae to embiids. However, frequency of emer-
gence of adults from my cultures suggests that their
larvae regularly utilize cover, and embiids as food, in
field colonies. Carrera (1947) reported a puparium
of the genus in embiid galleries in the botanical gar-
den in Rio de Janeiro.

ECTOPARASITOIDS: The most interesting ec-
tophagic parasitoids are larvae of wasps of the fami-
ly Sclerogibbidae (Chrysidoidea). The only recorded
hosts of this family are Embiidina. Male sclerogibbids
are black, slender, alate, |—4 mm in length. Females
(Fig. 39) are always apterous and move with great
swiftness within the galleries, however, like their
hosts, they spend much time resting motionless in
upper levels of the galleries spun in culture jars.

First instar larvae of sclerogibbids are legless, but
otherwise resemble meloid triungulins. They are scle-
rotic, well-segmented, and attach themselves to mem-
branes in various parts of the host’s body and, soon
after beginning to feed, become maggot-like. As a
rule, only one larva is attached to a host (Fig. 40), but
as many as four larvae of a large species have been
found attached to a single embiid (Fig. 41). Larvae
of certain small species of sclerogibbids may be more
numerous on a single host; a dozen or more may be
attached in a neat row on each side of the venter of
the abdomen.

When a sclerogibbid larva completes its feeding,
it drops off of the host and spins an elongate whitish

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 33

cocoon within a gallery, and pupates. Cocoons of
some species are coated with debris, which must have
been added by an embiid because they characteristi-
cally eject or cover foreign objects encountered in
their galleries. The cocoon is attached to the gallery
wall near the shrivelled, dead body of the host.

The sclerogibbids were revised by Richards (1939)
and there have been more recent papers dealing with
the systematics and biology of several species.
Scores of host-associated new species await descrip-
tion in California Academy of Sciences as a result of
my fieldwork and culturing activity. I have had no
success in continuous rearing of the wasps in labora-
tory cultures but this may be due to absence in a
culture of both sexes at the same time to insure mat-
ing. Shetlar (1973) succeeded in getting unmated fe-
males of Probethylus schwarzi to reproduce, but, of
course, the progeny were males. In another example,
fifteen males, but no females, emerged from a cul-
ture of Archembia n. sp., from Santa Catarina state,
Brazil.

FIGURE 39. Universal appearance of female sclerogibbids.
Note peculiar enlarged forelegs. Superficially, species dif-
fer only in size and coloration. Body length 3.0 mm.

Sclerogibbids are themselves parasitized by
wasps of the chalcidoid family Perilampidae, as evi-
denced by the emergence of five Perilampus
philembia Burks from as many cocoons of a sclerogib-
bid parasitoid of Archembia n. sp. in Tingo Maria,
Peru. More recently, apparently the same species
emerged in a large culture of Gibocercus n. sp. from
Ecuador’s Rio Napo region. A pupa of the Tingo
Maria sclerogibbid also yielded a male chalcidoid of
the genus Mondontemerus (Torymidae).

It is possible that sclerogibbids may be attacked
by their hosts. Three large males collected in a cul-
ture of Neorhagadochir Ross from Nicaragua suf-
fered extensive loss of antennal segments and this

FIGURE 40. Single sclerogibbid wasp larva feeding on adult
female of a new genus and species of Anisembiidae occur-
ring in upper Amazon of Brazil and Peru.

SS
Re E
Re
&

FiGuRE 41. Mature sclerogibid larvae attached to adult
female of Archembia n. sp. Tingo Maria, Peru.

34 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

could only have resulted from nibbling by the host
embiids.

Another type of ectophagy involves larvae of
small cecidomyid gnats of the genera Feltiella, or
Lestodiplosis (Ross 1944:491). These are similar to
sclerogibbid larvae both in appearance and method
of feeding. They were encountered on one occasion
only, in northern Florida, on Diradius vandykei
(Ross).

ENDOPARASITOIDS: Embiids, at least in
Neotropical regions, apparently are hosts to braconid
wasps of the genus Sericobracon (Doryctinae). Stud-
ies of such wasps were conducted by Scott R. Shaw
and Janice S. Edgerly (1985).

Embiidina are hosts to larvae of unusual tachinid
flies as evidenced by collections I have made in widely
separated geographic regions. Only two of the sev-
eral new species have been described. Rossimyiopsis
whiteheadi Mesnil, 1953, was reared from
Apterembia cercocyrta (Krauss) in South Africa and
E. I. Schlinger and I reared another series, Perumyia
embiaphaga Arnaud (1963), from Clothoda
longicauda Ross in Tingo Maria, Peru. Series I reared
from various embiid hosts occurring in Central Amer-
ica, Africa and tropical Asia await study. These ta-
chinids are small, averaging about 3 mm in length;
with shiny, not densely setose, black bodies, and usu-

= < der o
oe ap ee
FIGURE 42. Adult female of Gibocercus n. sp. from Ecuador resting on her mass of eggs
covered with layers of silk. Note egg parasite approaching from her rear.

ally smoky wings. As they walk they rotate their
wings.

EGG PARASITOIDS: Tiny wasps developing
within embiid eggs belong to genera Embidobia
Ashmead and Palaeogryon Masner, tribe Embidobi-
ini, family Scelionidae. They occur almost through-
out the range of Embiidina. The writer has reared
and preserved numerous host-associated lots of spec-
imens which should represent many new species. This
collection is being studied by L. Masner. The fe-
males of some species are pale ferrugineous and of-
ten apterous or subalate (brachypterous). Successive
generations can be reared in embiid cultures. At times
only one sex of the wasps, usually males, appear ina
particular culture.

The tendency of most embiid species to coat their
eggs with a hardened paste of chewed debris and their
feces, may reduce oviposition by the wasps. Guard-
ing by the parent female may also protect a large per-
centage of the eggs in a mass (Fig. 42).

PATHOLOGICAL HAZARDS: Disease epi-
demics may weaken and even kill all individuals ina
culture. It is assumed that such diseases also occur in
nature but probably do not have such a catastrophic
effect due to scattered occurrence of host colonies
and consequent reduced contagion.

ROSS: EMBIA, BIOSYSTEMATICS OF THE ORDER EMBIIDINA, PART 2 5

No studies specific to embiids have been made
of these diseases but one can speculate that they are
caused by viral, bacterial and fungal pathogens not
necessarily restricted to embiids. In some cases ac-
tivity of the victim will slow and eventually cease,
the body turning reddish as its tissues liquefy. In
cases of fungal infection, white mycelia begin to
outline somites and sclerites of the embiid victim
and later its entire body becomes a fuzzy, moldy mass.
Gerald M. Thomas of the University of California
identified the pathogen of one such epidemic, in a
culture of an oligotomid from northwest Thailand,
as Beauveria bassiana. He commented that this is
the most commonly occurring insect pathogenic fun-
gus in the world, and has a very wide host range on
terrestrial and aquatic arthropods.

The most important epidemics, however, are
caused by sporozoan parasites of the genera Adelina
(A. transita according to J. P. Kramer, pers. com.),
An infected culture

Gregarina, and Diplocystis.

exhibits gradually reduced vitality, no new galleries
are spun and eventually all the occupants die. Stefani
(1959, 1960) made special studies of Diplocystis
clerci parasitizing Haploembia solieri in the Medi-
terranean region. He noted that the protozoa may
damage sperm and weaken all males and thereby
limit a species’ reproduction to a residual minority
of females capable of reproducing parthenogeneti-
cally. Perhaps the consistently parthenogenetic form
of H. solieri developed this way in populations on
islands on the Tyrrhenean Sea on which all males
had been exterminated by repeated epidemics (see
section on parthenogenesis for details).

Undoubtedly, numerous other microorganisms
infect embiids throughout the order’s range; how-
ever, the only important investigations to date, are
limited to Sardinian hosts studied by Stefani who
also reported the parasitic nematode, Hexamermis,
in two species of Embia.

FIGURE 43. Culturing is a way of studying embiid biology and securing specimens for study. Those from Australia were
maintained in my home laboratory. Each (left) had an associated jar (right) in which “harvested” series were preserved

in alcohol jars.

FIGURE 44 Left, a gallon-sized culture (cover removed)
produced hundreds of adult specimens of Dihybocercus
n. sp. from Zambia. These half-grown, orange-and-black
nymphs mimic poisonous paederine staphylinid beetles.

36 OCCASIONAL PAPERS OF THE CALIFORNIA ACADEMY OF SCIENCES, NO. 149

Literature Cited

Arnaud, P. H. 1963. Perwmyia embiaphaga, a new
genus and species of Neotropical Tachinidae
(Diptera) parasitic on Embioptera. Amer. Mus.
Novitates No. 2143:1-8, 14 figs.

Carayon, J. 1974. Etudes sur les Hémiptéres
Plokiophilidae Ann. Soc. Ent. France (N.S.),
10(3):449-525, 33 figs.

Carrrera, M. 1947. Sobre 0 genero Leptopteromyia
Williston, 1908 (Diptera, Asilidae). Papeis
Avulsos do Dept. Zool. Sec. Agric., Sao Paulo,
8:89-96.

Edgerly, J.S. 1987a.
webspinner (Order Embiidina).
Entomol. 12;1-11.

Maternal behavior of a
Ecological

Edgerly, J. S. 1987b. Colony composition and some
costs and benefits of facultively communal be-
havior in Trinidadian webspinner, Clothoda
urichi (Embiidina: Clothodidae). Ann. Entomol.
Soc. Amer. 80:29-34, 6 figs.

Edgerly, J.S. 1988. Maternal behavior of a
webspinner (Order Embiidina): mother-nymph
association. Ecological Entomol. 13:263-272.

Edgerly, J. S. 1994. Is group living an antipredator
defense in a faculatively communal webspinner
(Embiidina: Clothodidae)? Jour. Insect Behav-
ior 7:135-147.

Friederichs, K. 1934. Des Gemeinschaftsleben der
Embiiden und Naheres zur Kenntnis der Arten.
Archiv fiir Naturg., N.F. 3:405—-444, 18 figs.

Mesnil, L. P. 1953. A new tachinid parasite of an
Embiopteran. Proc. R. Entomol. Soc. London
(B) 22:145-146, | fig., pl I.

Myers, J.G. 1928. The first known embiophile,
and a new Cuban embiid. Bull. Brooklyn
Entomol. Soc. 23:87—90, | fig.

Richards, O. W. 1939. The Bethylidae Subfamily
Sclerogibbinae (Hymenoptra). Proc. R. Entomol.
Soc. London (B) 8:211—223, 17 figs.

Ross, E. S. 1940. A revision of the Embioptera of
North America. Ann. Entomol. Soc. Amer.
33:629-676, 50 figs.

Ross, E. S. 1956. A new genus of Embioptera from
Baltic Amber. Mitt. Geol. Staatinst. Hamburg,
25:76-81, 2 figs.

Ross, E.S. 1957. The Embioptera of California.
Bull. California Insect Survey 6:51-57, 7 figs.

Ross, E.S. 1961. Parthenogenetic African
Embioptera. Wasmann Jour. Biol. 18:297-304.

Ross, E. S. 1966. A new species of Embioptera from
the Galapagos Islands. Proc. California Acad.
Sci. 34(12):499-504, 1 fig.

Shaw, S.R. and J.S. Edgerly. 1985. A new
brachonid genus (Hymenoptera) parasitizing
webspinners (Embiidina) in Trinidad. Psyche
92:505-511.

Shetlar, D. J. 1973. A redescription and biology of
Probethylus schwarzi Ashmead (Hymenoptera:
Sclerogibbidae) with notes on related species.
Entomol. News 84:205—210, 5 figs.

Stefani, R. 1953a. Un particolare modo di
accoppiamento negli Insetti Embiotteri.
Rediconti Accad. Nazionale, Lincei (8) 14:544—
549, | fig.

Stefani, R 1953b. La fisiologia dell’ accopiamento
in Haploembia solieri Ram. (Embioptera:
Oligotomidae). Rediconti Accad. Nazionale,
Lincei (8) 15:211-216, | fig.

Stefani, R. 1956. Il problema della partenogenesi
in Haploembia solieri Ramb. (Embioptera:
Oligotomidae) Atti Accad. Nazionale, Lincei (8)
5:127-201, 58 figs., pl. 1-2.

Stefani, R. 1959. Dati sulla partenogenesi
occidentali negli Embiotteri della fauna
mediterrane. Rendiconte Accad. Nazionale
Lincei, Classe Scienze fisiche matematiche
naturali (8)25:622-625.

Stefani, R. 1960. I rapporte tra parassitosi sterilita
maschile partenogenesi accidentale in popolazio
naturali di Haploembia_ solieri Ram.
Anfigonica. Revista Parassitologia 21:277-287,
4 figs.

Stefani, R. 1961. La citologia della partenogenesi
in due nuovi Embiotteri dell’ Africa tropicale.
Atti Accad. Nazionale Lincei (8) 30:254—256,

1 pl.

TT
3 9088 01302 6422